阅读说明：本技术 人类微生物组的衰老标记物和微生物组的衰老时钟 (Aging marker for human microbiome and aging clock for microbiome ) 是由 A·M·阿利佩尔 F·加尔金 A·扎沃龙科夫斯 于 2019-10-23 设计创作，主要内容包括：根据受试者的微生物丛的微生物分类剖析预测其表型年龄的方法可包括：从该受试者的微生物群系样本中分离出多个微生物核酸；分析该多个微生物核酸,以根据该多个微生物核酸鉴定出微生物群系的微生物数量。根据每种微生物的数量生成受试者微生物群系的分类学剖析；以配置有机器学习平台的计算机处理微生物群系的分类学剖析,以预测受试者的表型年龄；生成带有受试者预测表型年龄的报告；以及向受试者提供该报告。(A method of predicting the phenotypic age of a subject based on the microbiologic profiling of its microbiota may comprise: isolating a plurality of microbial nucleic acids from a microbiome sample of the subject; analyzing the plurality of microbial nucleic acids to identify a microbial population of the microbiome based on the plurality of microbial nucleic acids. Generating a taxonomic profile of the subject microbiome according to the number of each microorganism; processing taxonomic profiles of microbiome with a computer configured with a machine learning platform to predict phenotypic age of a subject; generating a report with the subject's predicted phenotypic age; and providing the report to the subject.)

1. A method of creating a biological clock of aging for a subject, the method comprising:

(a) receiving a biomarker signature from a sample of a subject;

(b) creating an input vector based on the biomarker features;

(c) inputting the input vector to a machine learning platform;

(d) generating a predicted biological aging clock for the subject according to the input vector of the machine learning platform, wherein the biological aging clock is specific to the subject; and

(e) a report is compiled including a biological aging clock and the predicted biological age of the subject is determined.

2. The method of claim 1, wherein the sample comprises another organism.

3. The method of claim 2, wherein another organism comprises at least one microorganism in the subject's microbiome.

4. The method of claim 3, wherein the sample is a stool sample of a subject.

5. A method of predicting a biological age of a subject from microbiota taxonomic profiling of a subject's microbiome, the method comprising:

obtaining a taxonomic profile of the subject's microbiome based on the number of each microorganism in the microbiome sample of the subject;

processing taxonomic profiles of the microbiome with a computer configured with a machine learning platform;

predicting a biological age of the subject from an output of the processed taxonomic profile; and

a report is generated with the predicted biological age of the subject.

6. The method of claim 5 further comprising:

receiving a plurality of microbial nucleic acids from a microbiome sample of a subject; and

the plurality of microbial nucleic acids is analyzed to identify the number of microorganisms in the microbiome based on the plurality of microbial nucleic acids.

7. The method of claim 6, further comprising generating a taxonomic profile of the subject's microbiome according to the number of each microorganism in the sample.

8. The method of claim 7, wherein taxonomic profiling comprises a particular microbiome.

9. The method of claim 8, wherein the particular microbial population comprises: a group 314; group 95; group 76; group 39; group 41; group 16; delaying aging population; a premature senility group; or a combination thereof.

10. The method of claim 9, wherein the specific microbiome comprises: group 39; or group 41.

11. The method of claim 5 wherein processing taxonomic profiles of microbiome results in the definition of one or more of: the relative number of total microorganisms in the microbiome; the relative number of each microorganism; and (or) the relative number of species of the microorganism and species-level taxonomic profiling.

12. The method of claim 11, further comprising generating a species-level taxonomic profile of the subject's microbiome based on the number of each microorganism.

13. The method of claim 12, wherein each taxonomic profile is based on whole metagenomic reads.

14. The method of claim 5 further comprising:

accessing a database having a plurality of reference microbiology profiles associated with biological ages of a plurality of reference subjects; and

the taxonomic profile of the microbiome of the subject is compared to a plurality of reference taxonomic profiles of microorganisms.

15. The method of claim 5 further comprising:

identifying an altered taxonomic profile that reduces the predicted biological age of the subject to a predicted biological age range that is lower than its predicted biological age; and

the altered taxonomic profile is included in a report provided to the subject.

16. The method of claim 5 further comprising:

identifying a processing method that results in an altered taxonomic profile of the subject to obtain a lower predicted biological age range for the subject; and

the processing method is provided in a report.

17. The method of claim 5 further comprising:

identifying at least one microorganism, the change in the amount of which provides an altered taxonomic profile to reduce the predicted biological age of the subject to within a predicted biological age range below its predicted biological age; and

providing information about the identified at least one microorganism in a report.

18. The method of claim 5 further comprising:

receiving a biological sample having a subject microbiome; and

isolating a plurality of microbial nucleic acids from a microbiome sample of a subject.

19. The method of claim 5 further comprising:

analyzing a plurality of microbial nucleic acids from a plurality of reference subjects, wherein each subject is analyzed for a plurality of microbial nucleic acids;

identifying the number of microorganisms in the microbiome of each reference subject based on the plurality of microbial nucleic acids;

generating a taxonomic profile of the microbiome for each reference subject based on the number of each microorganism;

processing taxonomic profiles of the microbiome of each reference subject with a computer configured with a machine learning platform to predict the biological age of each reference subject; and

the predicted biological age and associated taxonomic profile for each reference subject are saved to a reference database, the predicted biological age being correlated to the taxonomic profile of the microbiome of each reference subject.

20. The method of claim 19 further comprising:

generating a computer program product stored on a tangible, non-transitory storage device of a computer, the program product, when executed, enabling the computer to:

accessing a reference database;

comparing the taxonomic profile of the subject to a reference database;

providing information about at least one specific microorganism, the microorganism modulation correlated with a predicted biological age of the subject;

generating a report with the provided information; and

causing the report to be provided to the subject.

21. The method of claim 5 further comprising:

creating an input vector according to taxonomic parsing;

inputting the input vector into a machine learning platform;

generating a predicted biological senescence clock of the microbiome based on the input vector of the machine learning platform, wherein the biological senescence clock is specific to the microbiome; and

a report is written to include the biological aging clock and to determine the predicted biological age of the subject according to the microbiome.

22. The method of claim 5 further comprising:

providing an internet application to the subject's computer;

receiving input from an internet application to obtain a report of the subject;

associating the generated report with the subject; and

the generated report is provided to an internet application on the computer.

23. The method of claim 22 further comprising:

providing an internet application with an optional selection;

receiving a selected selection input from an internet application, the selected selection input obtained from a subject who selected the selectable selection;

determining information about the predicted biological age from the selected selection input; and

information about the predicted biological age is provided to an internet application on the subject's computer.

24. The method of claim 23 further comprising at least one of:

providing information about the type of microorganism in the microbiome;

providing dietary information on how to increase a particular microorganism of a microbiome;

providing dietary information on how to reduce specific microorganisms of a microbiome;

providing microbiome health information regarding a microbiome reference population within a defined age range;

providing information about reference subjects having similar microbiome taxonomic profiles; or

Providing a plurality of sequences of predicted biological ages of a subject;

wherein the internet application on the subject's computer is provided over the internet.

25. A method of creating a microbiomic aging clock, the method comprising:

obtaining microbiome nucleic acid information for a plurality of subjects;

identifying a microbial abundance profile for the microbiome based on the nucleic acid information for each subject;

training a plurality of neural network models with abundance curves;

evaluating performance of a plurality of trained neural network models;

determining a trained neural network model having a lower error threshold;

combining the determined trained neural network models into a set model; and

an aggregate model is provided.

26. The method of claim 25 further comprising:

filtering microbial lineage nucleic acid information prior to identifying the abundance curve;

filtering and normalizing the abundance curve; and

determining a cross-validated dataset;

wherein the training uses a cross-validation dataset with filtered and normalized abundance profiles.

27. A method of assessing the effect of dietary changes on a subject's predicted biological age, the method comprising:

obtaining a first biological age of the subject from a first microbiota taxonomic profile of a microbiome of the subject, the method comprising:

providing an indication to the subject to alter his diet by altering consumption of at least one substance;

obtaining a second biological age of the subject from a second microbiota taxonomic profile of the subject's microbiome;

comparing the first biological age to the second biological age by identifying a difference between the first microbiota taxonomic profile and the second microbiota taxonomic profile;

identifying at least one microorganism having an abundance that varies from a first microbiota taxonomy profile to a second microbiota taxonomy profile; and

providing a report that accounts for the difference between the first biological age and the second biological age and accounts for the change in abundance of the identified at least one microorganism.

28. The method of claim 27, wherein indicating comprises at least one of:

increasing uptake of a microbial substance whose increase is associated with a monotonically increasing effect;

decreasing the uptake of a microbial substance whose increase is associated with a monotonically decreasing effect;

increasing the uptake of a microbial substance whose decrease is associated with a monotonically decreasing effect; or

Reduction of uptake certain reductions of microbial material are associated with a monotonically increasing effect.

29. A method of predicting a subject's Body Mass Index (BMI), the method comprising:

obtaining a taxonomic profile of the subject microbiome based on the number of each microorganism in the subject microbiome sample;

processing taxonomic profiles of microbial communities with a computer configured with a machine learning platform trained using taxonomic profile data of known BMI values;

predicting the BMI value of the subject based on the processed output of the taxonomic profile;

generating a report using the predicted BMI value of the subject; and

providing the report to the subject.

30. The method of claim 29 further comprising:

creating an input vector based on the taxonomic profile and the associated BMI values;

inputting the input vector to a machine learning platform;

generating, by the machine learning platform, a predicted BMI clock for the microbiome based on the input vector, wherein the BMI clock is specific to the microbiome; and

a report is written to include the BMI clock and to determine a predicted BMI value for the subject based on the microbiome.

31. A method of predicting a disease state in a subject, the method comprising:

obtaining a taxonomic profile of the subject microbiome based on the number of each microorganism in the subject microbiome sample:

processing taxonomic profiling of microbial communities with a computer configured with a machine learning platform trained with taxonomic profiling data of known disease states;

predicting a disease state of the subject from the output of the processed taxonomic profile:

generating a report of the predicted disease state of the subject; and

providing the report to the subject.

32. The method of claim 31 further comprising

Creating an input vector based on the taxonomic profile and the associated known disease states;

inputting the input vector to a machine learning platform;

generating a predicted disease state clock for a microbiome from an input vector of a machine learning platform, wherein the disease state clock is specific to the microbiome; and

reports are written to include the disease status clock and to determine the predicted disease status of the subject according to the microbiome.

33. The method of claim 32, wherein the known disease state is type I diabetes.

34. A method of generating a composite taxonomy profile, the method comprising:

providing a microbioomic aging clock that has been trained with abundance curves of a microbiome for a plurality of subjects based on the nucleic acid information for each subject;

generating at least one composite taxonomic profile; and

for which at least one composite taxonomic analysis report is generated.

35. The method of claim 34 further comprising:

inputting criteria for synthetic subjects with defined phenotypes into a microbioomic aging clock; and

generating a synthetic taxonomic profile for the synthetic subject according to the determined phenotype.

Background

The age is widely accepted as defined by two; the first is defined as Chronological Age (CA), which refers only to the actual duration of a person's life. The second is Biological Age (BA), or physiological age, which is related to the actual physiological health of the individual. The biological age value is related to how the different organs of the body and the performance of the regulatory system are, and to what extent the overall balance of the body is maintained at all levels. Measurement of any physiological process in the body is typically accomplished by a set of predefined biomarkers. Biomarkers are a property that is measured as an indicator of normal biological processes, pathogenic processes, or pharmacological responses to therapeutic intervention. The goal of biomarker development is to measure very specific processes in vivo. However, aging is a systemic process and developing biomarkers for measuring biological aging in an individual faces particular challenges. In fact, such biomarkers must not only be objective and quantifiable substances, plus characteristics of the biological aging process that are easily measurable, but also be able to account for a set of changes that aging is not a single specific process, but is perceived in multiple physiological systems.

Chronological age is proportional to the risk of various adverse conditions and diseases, such as arthritis, heart failure, dementia, muscle weakness, etc. These associations are found in a number of studies throughout the population and can be used to facilitate specific healthcare guidelines and suggested interventions for a population. Age-based demographic studies are essential to officials who rely on such studies to set retirement and healthcare policies. However, the concept of chronological age tends to be impractical on an individual level. The life expectancy of a person may deviate considerably from the demographic estimates depending on the situation.

The biological age concept of an organism suggests that numerous cues may be aggregated into a single numerical score to more accurately predict risk and life expectancy of age-related disease (AAD) than chronological age. In a sense, chronological age and biological age are opposite concepts. The risk of age-related diseases in the population is a function of chronological age, while biological age is a function of one of such risks. To illustrate this concept, people often try to infer the artificial age or youth in front of the eye by observing their gait, wrinkles, skin pigments and voice when estimating the biological age. On the other hand, determining chronological age requires verification of the person's official documentation (e.g., birth certificate, etc.).

Heretofore, several attempts have been made by scholars to develop a clock for measuring biological aging. Both gene expression (Wolters and Schumacher, 2013) and DNA methylation profiling (Horvath, 2013) produce changes during senescence and can serve as biomarkers of senescence, as previously demonstrated by epigenetic clocks. Many studies analyzing transcriptomes of living tissues of various diseases have shown that patient age and gender have a significant effect on gene expression (Chovers et al, 2003), and that gene expression in mice has a significant change with increasing age (Weinruch et al, 2002) (Park et al, 2009), thus developing a database of mouse aging gene expression (Zahn et al, 2007) and for humans (Blalck et al, 2003) (Welle et al, 2003) (Park and Prolla, 2005) (Hong et al, 2008) (de et al) (de 2008)Curdo and Church, 2009). Despite the numerous aging biomarkers, the same individual microbiota may vary greatly from stage to stage and from person to personThe development of gut microbiome-based biomarkers has presented challenges to researchers.

The human gut is deposited by a dense microbial community, on a scale of 10¹⁴Granulosa cells, an order of magnitude higher than the number of cells in the human body (Suau et al 1999). The intestinal microflora refers to microorganisms in a specific site (e.g., intestinal tract), and is a complicated ecosystem that carries a variety of important functions in the organism. In addition to being a core element of the digestive system, the microbiota has the functions of regulating immunity, treating foreign organisms, producing important metabolites and even affecting higher nerve functions (De Palma et al, 2017; Rowland et al, 2018; Wu and Wu, 2012).

However, this effect is not unilateral. The microbiome is not solely responsible for the characteristics of certain hosts, but continuously responds to signals from the hosts through various feedback loops (Lozupone et al, 2012). Some of these feedback loops have been found to be reflected in the composition of the microbial community. Although this example shows inheritance of a distinct deleterious microflora, the gut microbiota ecosystem typically changes throughout the life of an individual. Some of these changes reflect individual lifestyles, while others appear to be more prevalent to the overall population.

Metagenomic studies provide valuable insight into how the microbiota in the gut microbiota evolves with age. Such studies indicate that intestinal tract colonization occurs at the time of parturition, i.e. the bacteria present in the birth canal. "Pioneer microbiome" includes facultative aerobes (Escherichia coli and Enterococcus enterobacteriaceae) which are replaced by obligate aerobes (such as Bifidobacterium infantis) after lactation: (Isocratic, 2017). Weaning results in another community being switched to the more typical adult microbiome (Tanaka and Nakayama, 2017). These early stages of colonization are important because the normal infant intestinal microflora promotes the formation of intestinal mucus, prevents the proliferation of pathogens,and regulatory T cells (Buford, 2017).

Although the inheritance of infant microbiota has been thoroughly studied by scholars and can be used to assess risk for various health conditions, the transition to adult microbiota is unclear. More importantly, compositional changes due to geographic location, medical history, diet, and other factors make analysis of adult microbiome difficult to be as productive as that of infants. Age-related studies of the human microbiome have failed to lead to a direct theory of gut microbiota aging. Some studies have shown a decrease in gut microbiome biodiversity in the elderly: (Isocratic, 2017; hopkins and Macfarlane, 2002). However, not all data sets are so, and the microbiome of healthy elderly people may be as diverse as young people (Bian et al, 2017; Maffei et al, 2017). Other findings include changes in the number of a particular taxonomic group (e.g., changes in strain, species or genus or other taxonomic class) or changes in the percentage of one or more particular microorganisms in the aging microbiome (e.g., strain, species or genus or other taxonomic class) in the total microbiome. Such as Bacteroides, Bifidobacterium, Blauettia, Lactobacillus or Ruminococcus, show a decrease in elderly humans, while Clostridium, Escherichia, Streptococcus, Enterobacteriaceae have been shown to increase in elderly humans (O' Toole and Jeffery 2015; Woodmansey et al, 2004). However, these patterns are not strictly defined as findings, as different findings may vary significantly.

To date, biologists appear to have divided the gut microbiome into three temporal states: the division rules of the three microbial groups of children, adults and the elderly are not clear. Among these, the adult microbiome is the most uncertain term; it has no inheritance phase as defined by newborns and does not usually reflect the gradient deleterious processes typical of the elderly organism.

Therefore, it would be advantageous to design a biological clock model that can estimate the age of an organism based on the gut microbiome. Furthermore, it would be advantageous to use a biological clock model to predict the age of an organism based on the gut microbiome. Furthermore, it would be advantageous if there were an application platform that allowed users to obtain information based on their gut microbiome and the age of the identified organism.

Disclosure of Invention

In certain embodiments, predicting a chronological age (e.g., the predicted chronological age is a biological age) and/or a phenotypic age (e.g., the phenotypic age is a phenotypic age based on the subject microbiome), the microbiota taxonomic profiling based on the subject microbiome, may comprise: isolating a plurality of microbial nucleic acids from a microbiome sample of a subject; analyzing the plurality of microbial nucleic acids to identify a microbial population of the microbial population based on the plurality of microbial nucleic acids; generating a taxonomic profile of the microbiome of the subject based on the number of each microorganism; processing the taxonomic profile of the microbiome with a computer configured with a machine learning platform (e.g., the machine learning platform includes one or more deep neural networks) to predict the age and/or phenotypic age of the subject; generating a report with the predicted age of the subject and/or the phenotypic age; and providing a report to the subject.

The methods predict how much age a subject is based on the calculated phenotypic age of the subject's microbiome, which results in the calculated phenotypic age being used as the predicted chronological age of the subject; however, the calculated phenotypic age may be different from the actual chronological age measured from the birth date of the subject. The predicted chronological age, i.e., the phenotypic age, is referred to herein as the biological age of the subject. Thus, the microbiome of the subject predicts the phenotypic age, determined and defined as the predicted biological age of the subject. Thus, a microbiome-based biological aging clock can be used to obtain a predicted biological age of a subject. Therefore, reference to the predicted age or phenotypic age will be made through a study of the predicted biological age.

In certain embodiments, a method of predicting a biological age of a subject may comprise: obtaining a taxonomic profile of the subject microbiome based on the number of each microorganism in the subject microbiome sample; processing taxonomic profiles of microbiota with a computer configured with a machine learning platform; predicting a biological age of the subject based on the processed output of the taxonomic profile; generating a report with the subject's predicted biological age; and providing the report to the subject. In certain aspects, the method may comprise: receiving a plurality of microbial nucleic acids from a subject microbiome sample; and analyzing the plurality of microbial nucleic acids to determine the number of microbial populations based on the plurality of microbial nucleic acids. In certain aspects, the method can include generating a taxonomic profile of the subject's microbiome based on the number of each microorganism in the sample. In certain aspects, the taxonomic profile includes a particular microbial population as defined herein, such as a particular microbial population comprising: a class group 314; class 95; class group 76; group 39; group 41; class 16; delaying aging groups; the premature senility group; or combinations of the above. In certain aspects, a particular microorganism population comprises: group 39; or class 41. In certain aspects, the method may comprise: receiving a biological sample having a subject microbiome; and isolating a plurality of microbial nucleic acids from a sample of the subject microbiome.

In certain embodiments, the method can include causing a taxonomic profile of the microbiome to be processed in a manner defining one or more of: the relative number of total microorganisms in the microbiome; the relative amount of each microorganism; and (or) species-level taxonomic profiling of relative numbers of various microorganisms. In certain aspects, the method can include generating a species-level taxonomic profile of the subject microbiome based on the number of the various microorganisms. In certain aspects, each taxonomic profile is based on the entire metagenomic read.

In certain embodiments, the method may comprise: accessing a database having a plurality of reference microbiologic profiles associated with biological ages of a plurality of reference subjects; and comparing the taxonomic profile of the subject microbiome to a plurality of reference microbiology taxonomic profiles.

In certain aspects, the method may comprise: determining an altered taxonomic profile to reduce the predicted biological age of the subject to a predicted biological age range younger than the predicted biological age of the subject; and incorporating the altered taxonomic profile into a report provided to the subject. In certain aspects, the method may comprise: identifying a process for obtaining an altered taxonomic profile of the subject to obtain a lower predicted biological age range for the subject; and providing the processing method in a report. In certain aspects, the method comprises: identifying at least one microorganism whose quantitative change provides an altered taxonomic profile to reduce the predicted biological age of the subject to a range of predicted biological ages younger than the predicted biological age of the subject; and providing information about the determined at least one microorganism in a report.

In certain embodiments, the method may comprise: analyzing a plurality of microbial nucleic acids from a plurality of reference subjects, wherein each reference subject is analyzed for a plurality of microbial nucleic acids; identifying the number of microorganisms of the microbiome of each reference subject based on the plurality of microbial nucleic acids; a taxonomic profile of the microbiome was generated for each reference subject based on the number of individual microorganisms. Processing the microbiome taxonomic profile of each reference subject with a computer configured with a machine learning platform to predict the biological age of each reference subject; and storing the predicted biological age and the associated taxonomic profile of each reference subject in a reference database, the predicted biological age being associated with the taxonomic profile of the microbiome of each reference subject. In certain aspects, the method comprises: generating a computer program product stored on a tangible, non-transitory storage device of a computer, the computer program product, when executed, causing the computer to: accessing a reference database; comparing the taxonomic profile of the subject to a reference database; providing information regarding modulating at least one specific microorganism associated with a predicted biological age of a subject; generating a report with the provided information; and causing the report to be provided to the subject.

In certain embodiments, the method may comprise: creating an input vector based on the taxonomic profile; inputting the input vector to a machine learning platform; generating, by the machine learning platform, a predicted biological senescence clock of the microbiome based on the input vector, wherein the biological senescence clock is specific to the microbiome; and preparing a report to include the biological aging clock and determine a predicted biological age of the microbiome based subject.

In certain embodiments, the method may comprise: providing an internet application to a computer of a subject; receiving input from an internet application to obtain a subject report; associating the generated report with the subject; and providing the generated report to an internet application on the computer. In certain aspects, the method may comprise: providing an internet application with selectable options; receiving an alternative selection input from the internet application, the alternative selection input being derived from the subject selecting a desirable selection; identifying information about the predicted biological age from the advisable selection input; and providing information about the predicted biological age to an internet application on the subject's computer. In certain embodiments, the method may include at least one of: providing information about the type of microorganism in the microbiome; providing dietary information on how to increase specific microorganisms of the microbiome; providing dietary information on how to reduce specific microorganisms of the microbiome; providing a reference group of microbiome with health information about the microbiome for a defined age range; providing information about reference subjects having similar microbiome taxonomic profiles; or providing the subject with a plurality of sequences of predicted biological ages, wherein the providing is provided via the internet to an internet application on a computer of the subject.

In certain embodiments, a method of creating a microbiome aging clock may comprise: obtaining microbiome nucleic acid information for a plurality of subjects; determining an abundance profile of the microbiome based on the nucleic acid information of each subject; training a plurality of neural network models with abundance profiles; evaluating the performance of a plurality of trained neural network models; identifying a trained neural network model having an error below an error threshold; combining the identified trained neural network models into a single set model; and providing the set model. In certain aspects, the method may comprise: filtering microbial lineage nucleic acid information prior to determining an abundance profile; filtering and normalizing the abundance profile; and defining a cross-validation dataset; wherein the training uses a cross-validation dataset with filtered and normalized abundance profiles.

In certain embodiments, a method of assessing the effect of dietary changes on a subject's predicted biological age may comprise: obtaining a first biological age of the subject based on a first microbiota taxonomic profile of the subject's microbiome, the method comprising: providing instructions to the subject to alter the diet by altering the intake of at least one substance; obtaining a second biological age of the subject based on a second microbiota taxonomic profile of the subject's microbiome; comparing the first biological age to the second biological age by determining a difference between the first microbiota taxonomic profile and the second microbiota taxonomic profile; determining a change in abundance of at least one microorganism from a first microbiota taxonomy profile to a second microbiota taxonomy profile; and providing a report of the difference between the first biological age and the second biological age, and the determined change in abundance of the at least one microorganism. In certain aspects, the description includes at least one of: increasing the uptake of a substance that increases the microorganism associated with a monotonically increasing effect; increasing the uptake of microbial material associated with the monotonically decreasing effect; increasing the intake of microbial material that reduces the associated monotonically decreasing effect; or decreasing the intake of microbial material associated with a monotonically increasing effect.

In certain embodiments, a method of predicting a subject's Body Mass Index (BMI) may comprise: obtaining a taxonomic profile of the subject microbiome based on the number of each microorganism in the subject microbiome sample; processing taxonomic profiling of microbiota with a computer configured with a machine learning platform, the platform being trained using taxonomic profiling data of known BMI values; predicting the subject's BMI based on the output of the processed taxonomic profile; generating a report with the subject's predicted BMI value; and providing the report to the subject.

In certain embodiments, the method may comprise: creating an input vector based on the taxonomic profile and the associated BMI values; inputting the input vector to a machine learning platform; generating, by the machine learning platform, a predicted BMI value clock for the microbiome based on the input vector, wherein the BMI value clock is microbiome specific; and preparing a report to clock in the BMI values and identify predicted BMI values for the microbiome-based subject.

In certain embodiments, a method of predicting a disease state in a subject may comprise: obtaining a taxonomic profile of the subject microbiome based on the number of each microorganism in the subject microbiome sample; processing taxonomic profiling of microbial communities with a computer configured with a machine learning platform trained with taxonomic profiling data of known disease states; predicting a disease state of the subject based on the output of the processed taxonomic profile; generating a report with the predicted disease state of the subject; and providing the report to the subject. In certain aspects, the method may comprise: creating an input vector based on the taxonomic profile and the associated known disease states; inputting the input vector into a machine learning platform; generating, by the machine learning platform, a predicted disease state clock for the microbiome based on the input vector, wherein the disease state clock is specific to the microbiome; and writing a report to include the disease status clock and determine a predicted disease status for the microbiome based subject. In certain aspects, the known disease state is type I diabetes; however, it should be recognized that microbiome taxonomic profiling associated with disease states of other diseases can be used to train the model. Thus, the model can be trained for any type of disease using microbiota data associated with the disease type, after which the subject can provide their microbiota nucleic acid for analysis to determine whether they have a defined disease state.

In some embodiments, a method of generating a composite taxonomy profile may include: providing a microbiome aging clock that has been trained in microbiome abundance profiling of a plurality of subjects based on nucleic acid information for each subject; generating at least one synthesis taxonomy profile; and generating a report for the at least one taxonomy analysis. In certain aspects, the method may comprise: inputting criteria for synthetic subjects having a defined phenotype into a microbiome aging clock; and generating a synthetic taxonomic profile for the synthetic subject according to the defined phenotype.

The above summary is illustrative only and is not intended to be in any way limiting. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features will be apparent by reference to the drawings and by the following detailed description.

Drawings

The above and following information, as well as other features of the present disclosure claimed, will become more fully apparent from the following description and appended claims, taken in conjunction with the accompanying drawings. Understanding that these drawings depict only several embodiments in accordance with the disclosure and are, therefore, not to be considered limiting of its scope, the disclosure will be described with additional specificity and detail through use of the accompanying drawings.

FIG. 1 shows a comparison of predicted age and observed age in a DFS model.

FIGS. 2A-2B show the comparison of observed and predicted ages for each age group in the DFS model (FIG. 2A) and the XGB model (FIG. 2B).

Fig. 3 shows an embodiment of the operation for medical purposes.

FIG. 4 shows a comparison of the importance levels of 74 taxonomic groups of microorganisms assigned by the DFS and XGB age prediction models for all samples (0-90 years) and age groups 15-30, 30-45, 45-60, and 60-90.

Fig. 5 shows venn plots of the 100 most important microorganism groups for age prediction by two machine learning methods.

FIG. 6 illustrates a computer or computing system that may be used in the methods described herein.

FIG. 7 illustrates an embodiment of operations for insurance.

FIGS. 8A-8X show ALE plots of 24 features in the DFS model.

FIG. 8Y shows the calculated ALE values of FIGS. 8A-8X.

FIGS. 9A and 9B provide several illustrations using ALE maps.

FIGS. 10A and 10B show typical age distributions in samples.

FIG. 11 is a schematic diagram of an optimal configuration of a neural network for DFS application testing.

FIG. 12 shows calculated ALE values for another set of microorganisms.

FIG. 13 shows calculated ALE values for another set of microorganisms.

FIG. 14 shows a method for constructing a microbiome aging clock.

Figure 15 contains data showing that different subjects have positive and negative effects on the predicted age of the same bacteria by using a biological clock model.

Figure 16A shows the predicted mean change with transferred abundance of microorganisms in a sub-sample of subjects over 50 years of age.

Figure 16B shows the effect of 6 microorganisms on a sample from a subject over 50 years of age.

Fig. 17 shows a graph of Kernel Density Estimation (KDE) for BMI value prediction.

Fig. 18 shows the model performance of independent sample sets obtained from public studies deposited in ENA servers.

FIG. 19 shows training a reliable aging clock model and applying cumulative local effects (ALE) to the model to measure its response to changes in feature values.

Fig. 20 provides an example of monotonically increasing and monotonically decreasing behavior of the gut microbiota aging clock.

FIG. 21 shows the 41 microbial signatures that exhibited the most prominent amplitude in a defined monotonic state.

FIG. 22 illustrates an embodiment of an application program that may perform the methods described herein.

Fig. 23 includes an embodiment of 30 profiles showing profile ID, actual BMI value, predicted phenotypic age, and chronological age.

Fig. 24 includes a chart showing the age distribution of samples (N1165 donors) used for "model 2" training.

Detailed Description

In the following detailed description, in the drawings, like numerals generally identify like parts unless context dictates otherwise. The illustrative embodiments described in the detailed description, drawings, and claims are not meant to be limiting. Other embodiments may be utilized, and other changes may be made, without departing from the spirit or scope of the subject matter described herein. It will be readily understood that the aspects of the present disclosure, as generally described herein, and illustrated in the figures, can be arranged, substituted, combined, separated, and designed in a wide variety of different configurations, all of which are explicitly contemplated herein.

The present invention relates generally to metagenomic biomarkers of human biological aging, which may include microbiome and microbiota of its microbiome. More specifically, the present invention relates to biomarkers included in the gut microbiome profile of a subject; these biomarkers can be processed and analyzed by a computer (e.g., machine learning, artificial intelligence, deep learning, deep neural networks, etc.) to provide indicators of humans and estimated biological ages (e.g., estimates or predictions of chronological age with unknown true age) using a microbiome aging clock. More specifically, generating a microbiome aging clock for an individual based on a subject's gut microbiome profile or other microbiome location can provide various uses, as described herein, including predicting a chronological age or phenotypic age (e.g., biological age) of the subject's individual. As used herein, the predicted age is not the true age, but a calculated phenotypic age. The phenotypic age of a microbiome refers to the age of the organism presented by the subject's intestinal microbiome phenotype.

In certain embodiments, machine learning and deep learning techniques are used to estimate the biological age of a human or other subject by profiling of metagenomic microbiomes. The machine learning platform may include one or more deep neural networks.

In some aspects, the machine learning platform may include one or more generative confrontation networks. In some aspects, the machine learning platform includes a resistant autoencoder architecture. In some aspects, the machine learning platform will include a feature importance analysis for ranking gene, genomic or taxonomic profiles and collections thereof by their importance in age prediction.

In certain embodiments, the present invention can be used to obtain a calculated microbiome profile based on gut microbiome profiles or microbiome profiles predictive of age (BA) of organisms, and then to generate interventions against biological aging. Thus, the microbiome profile of the gut or other location may be profiled or may be used to determine interventions to reduce aging in individual subjects.

In certain embodiments, a method of estimating a predicted Chronological Age (CA) of a subject based on taxonomic analysis of the microbiota of the subject is provided. The method can assess the importance of a particular taxonomic group of microorganisms in the aging process of a subject. The true chronological age can be compared to a predicted value (e.g., biological age) to assess the health of the subject.

In certain embodiments, the method may take into account chronological age and biological age. This mixture of chronological and biological ages can be used in a chronograph in bioscience, such as the biological chronograph described herein. In certain aspects, the biological aging clock is a mathematical model that assimilates a series of aging biomarkers, such as one of the specific groups defined herein, and produces an estimate of chronological age. The biological age-clock model may be validated (e.g., trained) on healthy individuals and, once trained, may be applied to individuals to determine the biological age of each subject. For example, the biological age of the subject may be evaluated as a normal healthy subject, or may be evaluated as a subject who is expected to be of a different age by taking anti-aging interventions or suffering from age-related diseases. The biological aging clock model shows accurate prediction for healthy people and reacts to age-changing conditions, and is therefore considered as a reliable method for estimating the biological age of humans, for example, for scientific progress.

The description herein provides a method to identify the most important features of any type of biological data in terms of aging, with limited dimensions (e.g., transcriptome, proteome, biochemistry, epigenome, metagenome). Such a method can determine which features have the greatest impact on the estimation of biological age and use these dimensions to design biological geriatric tools and interventions.

Although the subject is typically a human, other mammals, such as poultry livestock (e.g., cattle, horses, chickens, turkeys, pigs, etc.), dogs, cats, and other species of animals, such as birds, reptiles, insects, etc., can also be used as the subject. Thus, the present invention can be used to screen individual subjects for their health status, or for the overall public health status of certain subjects, whether in a common location (e.g., a city), or on a common heritage or other linked basis. The identified biological age can then be used to provide a medical recommendation to the subject or population of subjects. Biological age can also be used to develop consumer products and services to beneficially affect the subject's intestinal microbiome. While microflora of the gut microbiota may be used, microflora of other microflora may also be used, and the location of such microflora may include the oral cavity, skin, eyes, ears, genitals, specific organs, or other locations other than the gut.

In some embodiments, the method of obtaining a sample may be a standard procedure. The sample may comprise an intestinal microbiome or a parenteral microbiome (e.g., genitourinary, skin, oral, or other list). Thereafter, the sample may be processed to obtain a profile of nucleic acids (e.g., any form of DNA, RNA, etc.). In addition, protein profiling can also be used and analyzed, since nucleic acids or proteins are typically analyzed as quantities or percentages or other value representations to describe the microbiota profile of a sample.

These samples may be obtained from any biological sample of the subject, whether taken from the exterior of the subject or from an interior region thereof. For example, a sample of the gut metagenome may be obtained by collecting stool or biopsy. Accordingly, blood samples, urine samples, sweat samples, hair samples, mucus samples, semen samples, vaginal secretion samples, or other body fluid or excreta samples may be used.

In certain embodiments, a method for predicting the biological age of a subject based on the taxonomic analysis of his microbiome is provided. The method can include (a) obtaining or isolating microbial nucleic acids (e.g., any form of DNA, RNA, etc.) from a microbiome sample of metagenomics of a subject; (b) determining or estimating the relative or absolute number or percentage or other value (e.g., by using whole genome sequencing WGS techniques) of a plurality of different microorganisms (e.g., species, or genus-level taxonomic group categories) in a sample to produce a taxonomic profile of the microorganisms, such as the microbiome of a particular microbiome (e.g., gut, mouth, skin, eye, ear, genitalia, urethra, anus, colon, gut, vagina, a particular organ, or other location); and (c) processing the taxonomic profile with a machine learning platform to predict the age of the subject (e.g., sample donor). The predicted age may be a biological age of the subject based on a microbiome taxonomy profile processed by the machine learning platform. In certain aspects, the determined phenotype or biological age may be assigned to the subject. In certain aspects, a sample can be processed and nucleic acid information obtained. The method may operate by receiving nucleic acid information and performing the calculations described herein.

The present techniques and protocols are generally described for a single subject; however, these protocols may be applied to multiple subjects, and may include simultaneously evaluating or otherwise analyzing multiple taxonomic or metagenomic profiles of multiple subjects. This may include comparing profiles of different subjects to each other and/or to standard profiles (e.g., having (1) a biological age that is lower than the chronological age, (2) a biological age that is the same as the chronological age, and/or (3) a biological age that is higher than the chronological age). The resulting data may then be included in reports reported to the subject, which may include determining a suggested diet plan, treatment plan, lifestyle plan, exercise plan, or other plan. A subject who desires to reduce the biological age or need thereof may thereafter utilize one or more programs that report thereto.

In certain embodiments, the protocol can include creating a metagenomic biological clock and using it to predict the biological age of the subject. The data associated with metagenomic taxonomic profiling based on genetic analysis associated with chronological age of the subject may be used to train a computer model (e.g., machine learning and deep learning techniques with one or more deep neural networks). The data may be input into such a computer model to train the computer model to correlate the data with age, thereby enabling the data to be received and correlated with chronological age and/or phenotypic age. For example, the model may be trained as follows: when a certain data set is used for analysis, the model may provide an output of a predicted age based on biomarkers and taxonomic profiles that match certain ages. The model outputs a predicted phenotypic age based on taxonomic profiling. Metagenomic biosciences can also be used to make healthcare guidelines that can be based on the collective data of taxonomic profiling of multiple subjects. For example, the scenario may include loading hundreds to thousands of metagenomic taxonomic profiles into the model to obtain information about the general condition of a person, whether the person is older than expected age (e.g., older than chronological age). This data can then be studied to determine whether certain diets result in a phenotypic age above the chronological age. Thus, the data can be analyzed to determine whether reducing dietary sugar can make a subject phenotypically younger (i.e., biologically younger).

In certain embodiments, the different microorganisms may include two or more of the following microorganisms, wherein each microorganism is defined herein by the microorganism number provided in parentheses: bacteria vulgatus (1), Fusobacterium Ulcerans (2), bacteria ovatus (3), Bifidobacterium bifidum (4), Chryseobacterium gallinarum (5), [ Eubacterium ] recipient (6), [ Eubacterium ] galli (7), Bifidobacterium longum (8), Alisterides finegol (9), Faecalacterium praerupti (10), [ Clostridium ] saccharolyticum (11), Ornitobacterium rhizoticum (12), Bacillus dorei (13), Paramonas (14), Lactobaccos (15), Bifidobacterium adolescentis (16), Bifidobacterium lactis (19), Bacillus mucilaginosus (32), Lactobacillus (23), Lactobacillus (32), Lactobacillus (23) Bacillus calceuris (37), Eggerthella lenta (38), Aspergillus sphanchnicus (39), Bacillus fragilis (40), Shigella sp.PAMC 28760(41), Rhodococcus sp.YL-1(42), Acidococcus sensors (43), Campylobacter jejuni (44), Streptococcus paraguarinis (45), Bifidobacterium angulus (46), Negaticoccus maliens (47), Veillella paravulus (48), Streptococcus salivarius (49), Streptococcus mangostis (50), Victillares 44730(51), Paracoccus sp.06), Streptomyces strain (52), Bacillus faecalis (55), Bacillus faecalis (70), Bacillus strain (60), Bacillus strain (71), Bacillus strain (70), Bacillus strain (60), Bacillus strain (70), Bacillus strain (71), Bacillus strain (55), Bacillus strain (60), Bacillus strain (71), Bacillus strain (55), Bacillus strain (23), strain (60), strain (strain) and strain (60) Erysipeliococcus bacterium I46(73), Commononas kersterii (74), Enterococcus faecalis (75), Coriobacteriaceae bacterium 68-1-3(76), Bifidobacterium breve (77), Collinella aeroginens (78), Mordavella sp.Marseille-P3756(79), Bacillus helcogenes (80), Vortella melanogenin (81), Lactobacillus ruminis (82), Rothcia mucor (83), Clostridium sp.H121(84), Klebsiella sp.2N3(85), Hafnia alvarezii (86), Clostridium kernarerium (87), Gothicillinaceae (88), Klebsiella sp.88 (88), Clostridium (89), Clostridium (94), Bacillus subtilis (94), Lactobacillus (7992), Lactobacillus casei (93), Lactobacillus casei (93) and combinations thereof. In certain aspects, the microorganisms of this paragraph are of a particular group and are used as a particular combination, and "group 95" herein is 95 particular members of that group. Accordingly, each of the 95 different types of microorganisms listed above are present in the group 95.

In certain embodiments, the method may comprise a microbiome or biomarker panel comprising the following groups: bacillus vulgaris (1), Commamonas kersteri (2), Bifidobacterium bifidum (3), Rhodococcus sp.yl-1(4), Aliskites finegol (5), Bacillus ovatus (6), Chryseobacterium gallinarum (7), Desufovibrio farringensis (8), Campybacter jejuni (9), Bifidobacterium longum (10), Aspergillus lanchoeus (11), Bacillus calcei (12), Aspergillus formigenes (13), Streptococcus faecalis (14), Ornithia chrysogenum (15), Bacillus faecalis (15), Paracoccus sp.06, Lactobacillus sp.16, Bacillus faecalis (19), Bacillus strain (31), Bacillus strain (19), Bacillus strain (31), Bacillus strain (32), Bacillus strain (31), Bacillus strain (19), Bacillus strain (31), Bacillus strain (31.31), Bacillus strain (23), Bacillus strain (32), Bacillus strain (23) [ Clostridium ] boltea (37), [ Clostridium ] saccharolyticum (38) and [ Eubacterium ] villii (39). In certain aspects, the microorganisms of the present paragraph are of a particular group and are used in a particular combination, where "group 39" designates 39 particular members of the group. Accordingly, each of the 39 mentioned different types of microorganisms are present in the group 39.

In certain embodiments, the set of specific features derived from the gut metagenomic information includes the relative abundances of the following microbial species: bacillus dots (1), Akkermanophila (2), Bacillus cellulolyticus (3), Shigella sp. PAMC 28760(4), [ Eubacterium ] recipient (5), Ruminococcus bicirtus (6), Methanobacterium smith (7), [ Eubacterium ] eligins (8), Bifidobacterium bifidum (9), Bacillus vulgatus (10), Bifidobacterium adolescents (11), Bacillus thetaiobacterium (12), Bacillus ovatus (13), Bacillus calccus (14), Bacillus calceus (15), Bacillus faecalis (16), Bacillus subtilis (18), Bacillus faecalis (19), Bacillus faecalis (20), Bacillus faecalis (23), Bacillus (32), Bacillus (23), Bacillus (18), Bacillus (23), Bacillus (31), Bacillus (23) Chryseobacterium gallinarum (36), Eggerthella lenta (37), Faecalibacterium prausnitzii (38), Lactobacillus amylophilus (39), [ Clostridium ] saccharolyticum (40) and Bacillus fragilis (41). In certain aspects, the microorganisms of this paragraph are of a particular group and are used in a particular combination, where "group 41" designates 41 particular members of the group. Accordingly, each of the 41 different types of microorganisms mentioned are present in the group 41.

FIG. 21 shows 41 microbial signatures; these features show the most prominent amplitude and definite state of aging. A deep training network (DNN) model is trained to predict the age of a donor based on its microbiologic classification profile. ALE analysis (described herein) showed that 41 features in the model can change the predicted age by more than 0.1 years. Of these 41 features, 9 were monotonically increasing (associated with increasing predicted age, the lightest part), 26 were monotonically decreasing (associated with decreasing predicted age, the darkest part), and 6 did not show monotonic behavior (e.g., hatched part). This 41 taxonomic groups can in certain embodiments serve as biomarkers of aging to reduce the cost and complexity of analysis of all taxonomic microbiome. In addition, it should be recognized that other groups of microbial populations may be similarly analyzed and used.

These microbial species can be considered as biomarkers of human age to develop hypotheses about the nature of aging or to design consumables (such as probiotics and prebiotics) to affect aging and diagnostic systems using this range of characteristics.

In certain embodiments, two or more microorganisms are used. In certain aspects, only an even number of microorganisms are used. In certain aspects, only an odd number of microorganisms are used. In certain aspects, five or more microorganisms are used. In certain aspects, only the microorganism of fig. 4 is used. In certain aspects, only the microorganisms of FIGS. 8A-8Y are used. In certain aspects, only the microorganisms of FIGS. 9A-9B are used.

Using ALE plots, such as the graphs herein, the effect of a particular microorganism on the predicted age within a model can be estimated. Microorganisms whose effects monotonically increase with abundance are referred to herein as "premature senescence", and microorganisms whose effects monotonically decrease are referred to as "senescence delaying" (see FIGS. 8A-8Y). Other information is provided herein. In some cases, an ALE is performed from a profile derived from a real profile, where the derived profile has certain changes.

Within this definition, the following species have the ability to delay senescence: acidaminococcus intestine; bacteroides caccae; bacteroides caecimuris; bifidobacterium catenulatum; bifidobacterium pseudostellatum; chryseobacterium gallinarum; desulfovibrio fairfieldensis; dialister pneumosintes; lachnocrossdium sp.YL32; lactobacillus amylovorus; megasphaera elsdenii; ornithobacter rhizothrahale; oxalobacter formigenes; prevotella jejuni; rhodococcus sp.YL-1; alisipes finegoldii; anaerostipes hadrus; bacteroides dorei; bacteroides ovatus; bacteroides vulgatus; a Bifidobacterium bifidum; a Bifidobacterium breve; bifidobacterium longum; blautia hansenii; flavonifractor planutii; methanobrevibacter smithii; odoribacter splanchnicus; (ii) Parabacteroides detasonis; prevotella intermedia, and the like. In certain aspects, the microorganisms of this paragraph are of a particular group and are used in particular combinations, where the "senescence-delaying group" designates a particular member of the group having a senescence-delaying effect. Accordingly, each of the different types of microorganisms mentioned are present in the senescence-delaying group.

Within this definition, the following species belong to the premature senescence type: acidococcus fermentans; a Bifidobacterium denum; enterococcus faecalis; faecalibacterium praussnitzii; hafnia sp.cba 7124; lactococcus lactis; parvimonas micra; (ii) a Pseudomonas aeruginosa; eggerthella lenta; escherichia coli; roseburia hominis; ruminococcus bicirculans; veillonella parvula et al. In certain aspects, the microorganisms of the present paragraph are of a particular class and are used in particular combinations, where "the premature aging class" designates a particular member of the class having the property of premature aging. Accordingly, each of the different types of microorganisms mentioned are present in the premature aging group.

The methods described herein use biological information, such as information about a microbial population thereof. Thus, the method includes obtaining biological information. The step of obtaining biological information may include obtaining a biological sample and processing the physical behavior of the sample (as described herein or generally known) to obtain the biological information. Accordingly, an embodiment of a method of obtaining biometric information may comprise: analyzing the plurality of microbial nucleic acids to determine the number of microbes in the host habitat; analyzing a plurality of nucleic acids of an organism to determine the number and location of genomic mutations; analyzing a plurality of nucleic acids of an organism to determine the intensity of gene transcription; analyzing a plurality of chemical compounds within the tissue of the organism to determine the concentration thereof; a plurality of cell types within the tissue of an organism are analyzed to determine their numbers. This may be done by obtaining information from a biological sample, or may provide data for biological information. Thereafter, the analysis action is performed.

In certain embodiments, the derived biomarkers are used to approximate the performance of an initial aging clock with a reduced number of features. In certain aspects, the reduced number of features is used to assess the health status of the organism. In some aspects, the reduced feature vector is stored in a database with regulated access. In some aspects, the machine learning model built on the reduced feature vectors may be used by professionals or the public to provide general lifestyle advice. In some aspects, machine learning models built on reduced feature vectors are incorporated into a collection with other machine learning models to improve their performance in age or health prediction.

In some embodiments, the monotonically increasing feature may be processed separately from the monotonically decreasing feature to provide two scores related to the age or health of the organism. Some embodiments may include only a monotonically increasing feature, while other embodiments may include only a monotonically decreasing feature. In certain aspects, the monotonic states of one or more particular features are identified on the same sample used for model training; in other aspects, the determination is made on an external sample. In certain aspects, the identified characteristic of monotonic behavior is used to design a consumable, dietary plan, intervention, treatment, or other service that delays aging.

In certain embodiments, a process having taxonomic profiling can include inputting and processing various profiles of single or multiple microorganisms with a computing system. For example, processing can include processing an absolute number or percentage of microorganisms (e.g., absolute abundance), genus level profiling (e.g., number or percentage of each microorganism in a defined genus), taxa-specific gene count of microorganisms (e.g., number or percentage of each microorganism in a defined taxa), ecogroup number or percentage (e.g., abundance), or other profiling information. The data may be processed in any manner (e.g., by a factor) to obtain processed data, which may be used for analysis and prediction. For example, a relative or absolute quantity may be multiplied by a certain coefficient and the result used in the method.

In certain embodiments, the method can include processing the biological sample to provide a taxonomic profile with the strain, species, or genus. This allows profiling to be specific to different strains, specific at the species level (e.g. including all organisms in a species) or specific at the genus level (e.g. including all microorganisms in a genus). Thus, the data may be processed to select a certain taxonomy for age prediction and/or to parse different taxonomies from each other to generate a separate profile for each separate taxonomy.

In certain embodiments, each taxonomic profile is based on 16S variable region data. In certain embodiments, each taxonomic profile is based on the entire metagenomic data.

In certain embodiments, the taxonomy can be considered a metagenomic profile. The metagenomic profile of the subject can be compared to a reference metagenomic profile or a plurality of reference metagenomic profiles. The reference metagenomic data may be from one subject or a combination of multiple subjects. The comparison can be used to further identify the biological age of the subject. In some cases, comparisons can be made to find similar microbial compositions that produce predictions of different ages, as well as various microbiota changes that can adjust or alter the predicted age. Such information can be included in a report, such as a report of suggested changes to the microbiota taxonomic profile of the subject.

In certain embodiments, the method may include obtaining personal information about the subject and incorporating the personal information into a process to determine the biological age thereof. The personal information may include personal medical history, family member's lifespan, type of disease or condition that the subject has or has still suffered, family member's disease or condition type, medical images of the subject and/or family member, medical diagnostic data of the subject and/or family member, current weight and/or weight history of the subject and/or family member, current body mass index of the subject and/or family member, subject and/or family member's biometrics, geographic location and/or occupancy history of the subject and/or family member's current occupancy, drinking or smoking habits, eating habits, exercise habits, other health or lifestyle related personal data. The family members may be one, two, three, four or five people within the blood. During the determination of biological age, the personal data can be compared to a subject's metagenomic profile or a reference metagenomic profile. In view of the personal data, this information can be used to explore the correlation between predicted biological age and taxonomically profiled data. The correlation of personal data with predicted age and classification profile data as performed in multiple profiles can be used to enhance prediction and to determine unhealthy personal data that may result in higher phenotypic ages, or to use healthy personal data to devise strategies (e.g., treatments, diets, exercise, lifestyle, etc.) to reduce the biological age of others.

In certain embodiments, the protocol can include specific analysis of the microbial taxa used to obtain the taxonomic profile, showing the relative or absolute number or abundance of genomic profiles for a particular taxa. Thus, the relative analysis may be of a particular taxa relative to other taxa in the subject profile, or relative to the same taxa in another subject or population of subjects, such as an average or average relative amount. The absolute number or abundance can be described as the number of microorganisms per taxa in metagenomic profiling.

In certain embodiments, while estimating or determining the number or percentage of each taxa, the protocol for obtaining a taxonomic profile can comprise generating a corresponding functional profile of the subject microbiome. In certain aspects, the protocol may use taxonomic profiling, which may include certain characteristics, i.e., bacterial species. The protocol can also use the same data to generate functional profiles, one characteristic of a gene family, which may contribute to: (1) the phenotypic age is lower than the chronological age; (2) the phenotypic age is the same as the chronological age; and (or) (3) the phenotypic age is higher than the chronological age. These methods can use functional profiling, modified to accept gene number, abundance, or percentage. The phenotypic age is the biological age of the calculated biological clock model.

In certain embodiments, the protocols described herein may include using the generated age prediction as a measure of phenotypic age. Thus, the phenotypic age may be compared to the true age of the subject. When there is a difference, e.g., the phenotypic age is higher than the chronological age, an intervening regimen may be implemented in an attempt to reduce the phenotypic age to the chronological age. This may include certain treatments, such as anti-aging treatments; the methods of treatment described in the incorporated references may also be used. When the phenotypic age is below the chronological age, the subject's genetic characteristics, proteomic characteristics, health characteristics, family health characteristics, eating habits, exercise habits, sleep habits, work habits, and other characteristics of the subject's life can be analyzed to determine parameters that may result in an age below the phenotypic age (e.g., more desirable) or above the phenotypic age (e.g., less desirable). These parameters can then be aggregated and provided to subjects with a phenotypic age higher than the chronological age so that they can attempt to apply these parameters to reduce the chronological age of the phenotype to drive it toward the chronological age. Thus, healthy subjects with younger phenotypic ages may be modeled as a model of health and behavior patterns so that their health and behavior patterns may be applied to other subjects in an attempt to reduce the phenotypic age toward or below the chronological age.

In certain embodiments, the protocol can be performed on several subjects such that multiple taxonomic profiles can be obtained and compared to each other, which also allows the multiple taxonomic profiles to be used to identify single or multiple taxa present in subjects having the following characteristics: (1) phenotypic age is lower than chronological age; (2) the phenotypic age is the same as the chronological age; and (or) (3) the phenotypic age is greater than the chronological age. Taxa involving any class can be determined throughout the subject population and used as criteria for determining the predicted phenotypic age. Thus, it can be determined that a change in the amount or percentage or other value (e.g., abundance) of at least one microorganism or microbiome can affect the predicted age of the subject. The identity of one or more microorganisms can be used to identify microorganisms useful for obtaining a lower phenotypic age, such microorganisms can be included in the treatment to provide such microorganisms to a subject having the same or a higher phenotypic age as the chronological age. Thus, beneficial microorganisms attributed to a lower phenotypic age are useful in therapy. On the other hand, microorganisms identified as contributing to a higher phenotypic age may be used to assess the phenotypic age of other subjects, as well as identified by means of elimination. That is, strategies can be used to remove such microorganisms from the subject's microbiome. For example, large scale cleanup of microorganisms can be performed followed by establishment of beneficial microorganisms associated with lower phenotypic age in the subject's microbiome. In addition, beneficial taxonomic profiling can be used to establish beneficial groups of viable microorganisms to establish a subject microbiome. If the phenotypic age is higher than the chronological age, information about the taxonomic profile and microorganisms therein and the treatment plan may be reported to the subject.

In some embodiments, multiple taxonomic profiles may be compiled in a single database, which may be aggregated or parsed into a database with profiles: (1) the phenotypic age is lower than the chronological age; (2) the phenotypic age is the same as the chronological age; and (or) (3) the phenotypic age is higher than the chronological age. Taxonomic profiling can also be resolved into categories of various diseases or disorders, such as age-related diseases or disorders (e.g., phenotypically age-related diseases or disorders). Thus, a regimen for a particular subject may include comparing the taxonomic profile of the subject to profiles in a database. This can be used to compare profiles of particular subjects to a reference database, and then assess the comparative phenotypic age of the subjects and/or risk of having age-related disease. Risk can be calculated as the probability of having a certain taxonomic group feature or taxonomic group feature associated with a disease or disorder. Information regarding the comparative phenotypic age and/or risk of developing age-related disease may be compiled into a report and this information may be reported to the subject.

In certain embodiments, the protocol can include establishing a database of reference profiles and predictions of the response of those profiles, i.e., chronological age, phenotypic age, or disease or disorder status associated with differences between true age and identified phenotypic age. The database may also include the health status of each subject, as well as information (e.g., personal data) about their various diets, exercises, lifestyle, or other behaviors.

In certain embodiments, the protocol can include establishing software that can be used to compare a taxonomic profile of a microorganism of a subject to a reference database. This may include creating software that can receive the subject's microbiology data and compare it to the microbiology data of one or more other subjects. The data on comparability and similarity can be used to determine the age of the reference or its average and use this age or average age to predict the phenotypic age of the subject. The software may also be used to provide information about the effect of a particular microorganism on the age prediction of an individual, and this information may be compiled into a report and reported to the subject.

In some cases, a database (e.g., a database with taxonomic profiling of microorganisms), whether or not associated with phenotypic age, may be configured as a blockchain storage system. A blockchain storage system can be used to track the progress (e.g., progression over time) of microbiologic profiling of one or more subjects. A blockchain storage system may be used to track predicted phenotypic age in view of microbial taxonomic profiling progress.

In some cases, a protocol may include training an age prediction clock (e.g., model, machine learning platform) only at temporal ages with or without associated taxonomic parsing. In addition, the protocol may translate methods to make a phenotypic aging clock in the presence of more personal data. Phenotypic age may be provided as a measure to a subject to inform the subject of the risk of having an age-related disease. In some cases, the protocol does not evaluate chronological age, only phenotypic age, and vice versa.

In certain embodiments, the database and report may contain information about at least one specific microorganism that can adjust age prediction. The microorganism can increase or decrease the predictive value of age. Such one or more microorganisms may be identified for analysis of taxonomic analysis to determine a predicted age. The presence, absence, or quantity of such one or more microorganisms can facilitate linking the taxonomic profile of the subject to phenotypic age estimates or predictors.

In certain embodiments, the regimen may comprise determining a diet plan for the subject that may result in a phenotypic age that is less than or equal to the chronological age. In certain aspects, the phenotype age-reduced dietary plan may be from another phenotype age lower than the current subject's dietary plan. A diet plan for a subject with a lower phenotypic age may be applied to a subject with a higher phenotypic age in an attempt to reduce the phenotypic age of the subject. In some cases, diet plans for a plurality of subjects with a lower phenotypic age may be compiled and analyzed to determine characteristics that are consistent among a high proportion of subjects with a lower phenotypic age, and then characteristics of such reduced phenotypic diet plans may be determined. Thereafter, the characteristics of the ascertained phenotype-reduced diet program can be applied to subjects with higher phenotypic ages. Thus, a productive diet plan can be used for the course of treatment of other subjects in need thereof.

In certain aspects, the regimen may include establishing a diet plan for supporting or inhibiting a particular microorganism whose change in number, percentage, or abundance may affect the predicted age, such as reducing the phenotypic age. After the dietary plan is created, it may be included in a report and provided to subjects in need of a reduced phenotypic age. In addition, the created dietary plan can be implemented as a therapeutic dietary plan for subjects in need of reduced phenotypic age.

In some embodiments, the method may comprise generating a suggested therapy for the subject based on its taxonomic profile. This can include determining metagenomic profiling and/or age prediction (e.g., phenotypic age), which can then be used to determine a suggested therapy. Suggested therapies may include traditional drug therapies, such as by administration of therapeutic agents (e.g., drugs, vaccines, etc.), as well as therapies that change the subject's microbiome to a healthy microbiome, which have been identified as being associated with a young phenotypic age. The alteration of the microbiome may include washing the microbiome to remove certain microbes or all microbes and supplementing the microbiome with selected microbes or taxonomic microbial profiling related to the age of the young phenotype. The protocol can include comparing the subject's metagenomic profile and age prediction to a reference metagenomic profile, determining a therapy, generating a report regarding the determined therapy, and providing the report to the subject.

In certain embodiments, a protocol can include analyzing information regarding the effect of at least one particular microorganism on age prediction. Thereafter, the regimen may include creating a microbiota therapy, e.g., to cause the subject's microbiota to tend to resemble healthy microbiota of a person with a young phenotype. To change the age prediction over time, microbiome therapy may be administered. Thus, the identified microorganisms can be used to develop and provide microbiota therapy to one subject as well as to multiple subjects in a population. A deep signature synthesis (DFS) model or extreme gradient boost (XGB) model may be used to define beneficial microorganisms (e.g., resulting in a younger phenotypic age) and/or detrimental microorganisms (e.g., resulting in a younger phenotypic age). These beneficial microorganisms can be included in a therapy or treatment applied to a subject to increase the number, abundance, or percentage of such beneficial microorganisms. Thereafter, the harmful microorganisms may be targeted for selective removal, or general microbial removal may be performed when the harmful microorganisms are found. Thereafter, the beneficial microorganisms can be used to restore the microbial flora to a beneficial state.

In certain embodiments, consumer products can be designed based on taxonomic profiling and associated predicted phenotypic age. This approach allows the model to estimate the impact of certain microorganisms on the aging process. Some dietary products may be specifically designed to alter the composition of the intestinal microbiota in a desirable manner, such as to increase the predicted phenotypic age. The consumer product may be personalized or may be designed according to a general pattern of age progression of the microbiome in the population to achieve a lighter predicted phenotypic age. Microorganisms determined to be age-related to the older predicted phenotype may be specifically excluded from the consumable. Although the technology is described with emphasis on the human gut microbiome, the same methods disclosed herein can be applied to other human habitats or other types of animals to achieve similar results. These taxonomically profiled and predicted phenotypic ages can be used to design a large number of consumer products that are intended to alter the corresponding microbiome, where some examples of consumer products may include clothing, body odor blockers, underwear, toothpaste, and oral care products, among many others. In addition, non-human consumables may also be designed based on age prediction and derivation of a marker set. For example, such methods may also be applied to modify the feed additive of cattle.

The identified beneficial microorganisms and identified detrimental microorganisms can be used to develop consumer products, articles of manufacture, any substances, or any items that can be provided to a consumer, which can be used to increase beneficial microorganisms in a subject's microbiome. These microorganisms can also be used to reduce harmful microorganisms in the subject's microbiome. These microorganisms can also be used to establish beneficial taxonomic profiles of microorganisms in the microbiome of a subject. In part, designing a consumer product may be based on profiling information in a profile reference database, and delivering the product to the consumer for: increasing beneficial microorganisms in a subject's microbiome; reducing harmful microorganisms in a subject microbiome; or establishing a beneficial taxonomic profile of microorganisms in the subject's microbiome. Several examples of commercial products may be cosmetics, dietetic products, medical products, clothing, instruments, candy or the like. Commercial products may include beneficial microorganisms, health-promoting substances, substances on which beneficial microorganisms depend (e.g., vitamin A, B, C, D, etc.), or the like.

In certain embodiments, information regarding identified beneficial and detrimental microorganisms and taxonomic profiling of beneficial microorganisms can be used to create cosmetic products to support or inhibit specific microorganisms whose changes in quantity, percentage, abundance, or other parameters affect age prediction (e.g., make age prediction lower). The cosmetic product may be delivered to and used by the subject. The cosmetic may include makeup, mascara, foundation, lip gloss, eye shadow, eyeliner, lip gloss, foundation, blush, lipstick, nail polish, concealer, foundation, bronzer, highlighter, lip gloss, lip liner, shampoo, hair conditioner, hair styling combination, hair dye, gloss essence, ointment, moisturizer, cleanser, anti-aging cream, exfoliant, eyedrops, and blemish control composition. Acne control ingredients, make-up removers, lotions, astringents, oral care ingredients, night creams, skin sunscreens, sun creams, toothpaste, lip plumpers, body lotions, perfumes, body washes, antiperspirants, deodorants, soaps, depilatory ingredients, sun care ingredients, body scrubs, bath salts, body washes, aromatherapy, cellulite treatment, scar reduction agents, skin care powders, and the like.

In certain embodiments, information about identified beneficial and detrimental microorganisms and taxonomic profiling of beneficial microorganisms can be used to create a diet product intended to support or inhibit specific microorganisms whose changes in abundance will affect age prediction (e.g., make the age prediction value younger). The dietary product can be delivered to and used by a subject. The dietary product may be any edible item such as a food, beverage, supplement, powder, pill, capsule or other item.

In certain embodiments, information about certain beneficial and detrimental microorganisms and taxonomic profiling of beneficial microorganisms can be used to create medical products intended to support or inhibit specific microorganisms whose abundance changes will affect age prediction (e.g., make age prediction younger). The medical product may be delivered to and used by a subject. Medical products may include products for use inside the body of a subject (e.g., catheters, stents, filters, needles, implants, etc.) or for use outside the body of a subject (e.g., bandages, stents, plasters, etc.).

In certain embodiments, information about certain beneficial and detrimental microorganisms and beneficial microorganism taxonomies profiling can be used to create clothing intended to support or inhibit particular microorganisms whose changes in abundance will affect age prediction (e.g., make age prediction younger). Such clothing may be delivered to and used by the subject. The garment may be any article worn by the subject, such as underwear, bras, pants, shorts, shirts, hats, gloves, socks, or the like.

In some embodiments, the method of determining a premium rate may take into account a predicted phenotypic age and taxonomic profiling. The insurance method may allow the insurer to estimate a rate that is more accurate for representing the predicted phenotypic age. Placing the customer into a forecast of a age group (e.g., a higher forecasted age) than they actually are provides a signal that the contra may have hidden conditions that affect life expectancy or risk of disease, thereby affecting the predetermined premium rate. For example, by correlating gut microbiota with certain health problems and using age prediction as a general marker of human phenotypic age, insurance agents can assess the likelihood of future claims. For example, if a healthy 30 year old male wants to buy a life insurance but the microbiome belongs to the more typical 60 year old, the agent can look up the risk of having heart disease in the reference database of the elderly and take this risk into account by weighting this age prediction among other criteria. On the other hand, if data provided by a person of age 60 is predicted to be younger, the customer may be entitled to a discount.

In some embodiments, to obtain a discount on a policy, the applicant may be willing or required to provide a taxonomic analysis of its microorganisms, as well as a predicted phenotypic age determined by the present invention.

In certain embodiments, to track a subject's phenotypic age or its taxonomic profile, a blockchain solution may be implemented. An immutable and tamper-resistant blockchain database would allow insurance agents to easily request data and allow customers to control who can access their personal data.

In certain embodiments, the predicted phenotypic age may be used to determine a policy for the subject. Thus, the insurance method may be configured to consider the predicted phenotypic age in order to increase policy costs when the predicted phenotypic age is above a threshold or chronological age. Further, the insurance method may be configured to consider the predicted phenotypic age so as to reduce policy costs when the predicted phenotypic age is below a threshold or chronological age. Accordingly, the insurance method may comprise making a policy on the basis of analyzing the metagenomic microbiology profile of the subject in view of the metagenomic microbiology profile in the reference database. Insurance policies parsed at least in part based on metagenomic microbiology can be delivered to insurance clients. In some cases, the policy holder's rate of payment is influenced by age predictions parsed from their microbiology, as described herein. In some cases, the policy holder is required to provide the underwriter with the dynamics of their validated biotaxonomic profiles and corresponding age predictions for analysis.

In certain embodiments, a method for creating a predictive computer model that analyzes a taxonomic analysis of microorganisms in one or more subjects and provides a predicted phenotypic age may be performed. The protocol may include analyzing aggregated data (e.g., taxonomic profiling of microorganisms and/or predicted phenotypic age) to generate a generative model for creating silicon-based microbiome communities. The silicon-based microbiome community may be configured to be difficult to distinguish from a real reference database. Thus, the population of siliceous organism microorganisms can be used in place of the actual reference database. This allows the age prediction method to use simulated data, such as the silico-based microbiome community, rather than real data of real subjects.

In some embodiments, a method of generating a composite taxonomy profile may include: providing a microbiome aging clock that has been trained in profiling the abundance of microbiome of a plurality of subjects based on the nucleic acid information of each subject; generating at least one composite taxonomic profile; and generating a report for the at least one taxonomic analysis. In certain aspects, the method may comprise: inputting criteria into a microbial senescence clock of synthetic subjects having defined phenotypes; and generating a synthetic taxonomic profile for the synthetic subject based on the defined phenotype. For example, generative models can be used to create a synthetic microbiome synthetic (e.g., computer simulated) microbiome community or taxonomic profile. That is, once the microbial aging clock is derived, it can be used to generate synthetic microflora populations, or synthetic taxonomic profiling, for synthetic subjects that are not real subjects. This provides predictive information for non-real subjects. In one embodiment, the parameter may be input into a microbiome aging clock to obtain a synthetic microbiome community or taxonomy profile of the input parameter, where the input parameter may include information such as health, weight, gender, age, any disease, any injury, or the like. In one embodiment, the input may include information about a 21 year old male with diabetes, and the microbiome aging clock may provide a prediction, or an example of one or more synthetic microbiome communities or taxonomic analyses, for the 21 year old male with diabetes. Thus, based on input from non-real subjects, a synthetic microbiome community or taxonomic profile of synthesis can be generated.

In some embodiments, such as using predictive computer modeling schemes, a predicted phenotypic age may be generated. Such predicted age may be provided as a numerical value. Furthermore, such an age prediction may achieve an average absolute error over several years, months, and/or days when the age of a subject is predicted based on the taxonomic analysis of the subject's microorganisms. Such a prediction of phenotypic age may be useful for forensic, medical, and scientific purposes.

To prevent the disclosure of personal information, the subject may digitally sign their samples with an asymmetric algorithm using their personal signing key. This allows the key holder to prove the authenticity of its microbiological characteristics and the corresponding age prediction without having to mention its digital identity in storage. In addition, the expert facility may also sign the configuration file and the predictive hash value to verify that data collection and analysis is performed by them.

To decentralize data analysis, prevent data correlation and accumulation for expert facilities, WGS machines may be connected to decentralised computing systems. For example: some sequencers do sequencing without having any information about the subject, however they receive a digital signature of the subject and save the signature-sequencing association in memory. It feeds back the output and signature to a decentralized computing system where the data is shared by all its agents through quality control, sorting and prediction algorithms. Such a system allows secure remote computing over an uninvolved range, ensuring that the agent does not have access to the data it analyzes. A subject may need to share their public key and signature path to a distributed storage location with a computing system in advance in order to be the only person who can be granted access to the output. The computing system checks the signatures of its inputs and redirects the output to some designated location. As such, metagenomic profiling and prediction can only be used by subjects: the expert facility performs sequencing but does not receive any output, and the computational network agent is not designed to have access to the sequencing data, with the final output being stored at the location specified by the subject. To verify the validity of all processing steps, each participant in the pipeline leaves an individual signature on the output hash value.

In certain embodiments, after the subject receives the data, the counterparty may share with an entity (e.g., with the insurance policy provider) that will be able to verify the authenticity of the prediction by examining the signatures accumulated during the process.

The current model can be obtained by machine learning.

The present invention can adapt several machine learning methods to age prediction. The machine learning problem is typically a supervised regression, and may be model trained with bacterial taxa features as input variables.

The data used to train the machine learning algorithm can be modified to contain only reliable detection features (>1e-5 abundance). For example, the size of the data after removing the age values that are not present or missing may be about 3,058 records from about 621 people. Each record may contain 1,673 bacterial signatures. The target variables may include an age in the range of 7 to 90 years. For example, the age distribution can be described in FIGS. 10A-10B.

These models can be evaluated by various protocols.

The model may be trained by quintuple cross validation and grid or random search strategies to compensate for overfitting. Since the size of the data set is rather modest, no external test set is introduced. Instead, a given index may be derived as a prediction for 5 test sets per aliquot, resulting in a fully predicted input data set. All models can be evaluated using the same scoring method for fair comparison. In all evaluations, the following indices may be measured.

Determining the coefficient (R)²) Is defined as:

whileIs the mean age in the data set, C_iIs the taxonomic profile of sample i.

The mean absolute value (MAE) can be defined as:

and C_iIs a taxonomic profile from sample i.

The Pearson's correlation coefficient (r) may be defined as:

wherein N is the number of samples, andis the average predicted or observed age.

R²The index may be selected as a target index for performance comparison. The performance of the different models and their configurations may be compared across the entire prediction dataset according to target metrics.

In some embodiments, the model may be configured as a neural network configuration, which may be a Depth Feature Selection (DFS) model. Among the choices of different neural network models, DFS can be used in the methods described herein for at least two reasons: first, the DFS model aims to identify the most important features, which is necessary for feature interpretation. Secondly, the DFS model means a deep architecture, which shows the best effect on the prediction task of high-dimensional data. The main idea of the DFS model is to add a sparse one-to-one linear layer between the input layer and the first hidden layer of the MLP. The goal of the DFS model is defined as:

the structure of the final model can be derived by adjusting the following hyper-parameters: number of layers and number of neurons in each layer, type of nonlinearity (ReLU, prellu, ELU, leakyreu), learning rate (0.001, 0.005, 0.001). Adam can be used as an optimizer, and a 0.5 ratio drop (Dropout) technique can be used to reduce overfitting. For architectures with 2 and 3 hidden layers, batch normalization techniques may be employed.

In certain embodiments, a feature analysis may be performed. To determine bacterial taxa (e.g., more than one taxa) important for age prediction, current protocols can employ the following techniques: envelope feature importance (PFI) and cumulative local effects (ALE).

In some cases, a PFI process may be used. The method measures target indexes (R) before and after characteristic line interchange²) To calculate a relative importance value for the feature in the age prediction. The basic assumption is that the interchange of important features results in a reduction in prediction accuracy compared to the original prediction. On the other hand, random permutations of non-essential features should not affect the prediction to a large extent. Due to the bias that causes random permutations, the importance score for each feature can be averaged over the results of k permutations. All significance scores can be calculated for all five trained models, followed by averaging to avoid bias tendencies of overfitting.

Wherein S is_originalScoring the target index, S_shuffled(k) The predicted target index is a guide parameter for controlling the number of permutations of each feature. Target metric selectable R²And (6) measuring.

The ALE method may also be used. The ALE method predicts the average change in the value of a feature when it changes slightly by measuring the effect of a particular feature on age prediction. The ALE plots are assembled by first computing the Local Effect (LE) for each quartile (ALE equation 1):

(ALE equation 1);

wherein N is_QA number of samples between the qth and (Q-1) quanta of the target feature; c_jAn abundance vector for sample j belonging to the interval; f (C)_jQ) is the predicted age of the vector, where the Q-th magnitude value of the feature is replaced with the target feature value.

To generate ALE, the LEs are added (ALE equation 2):

ALE₁＝LE₁

ALE_Q＝LE_Q+ALE_Q-1

(ALE equation 2)

Furthermore, the ALE is centered, so the sum of all ALEs for a feature is zero.

In some embodiments, a gradient may be used to boost the classifier. To validate the selected set of labeled features, a gradient boosting classifier was trained based on 96 features (95 from the labeled set in table 5 and 96 th, equal to all other microbial abundances). The data set can be divided into a training set and a test set, the training set containing 75% of all individuals. All samples can be divided into three groups: young (15 to 40 years), middle aged (41 to 60 years) and elderly (61 to 90 years). The classifier can be implemented using the standard GradientBoosting Classifier class of Python SciKit, set forth below. { ' n _ estimators ':100, ' max _ depth ':4, ' min _ samples _ split ':4, ' learning _ rate ':0.2, ' loss ': device ' }. Table 3 shows the predicted quality index set one-to-one, and table 4 shows the quality of the stochastic classifier. In this simple setup, the disclosed marker list provides 25% more advantage than random.

In some embodiments, the method of predicting the age of an organism may include neural network, gradient boosting, random forest based models.

Fig. 1 shows the comparison of predicted age and observed age in the DFS model. Thus, FIG. 1 illustrates age prediction using an embodiment of the present invention. Using the Depth Feature Selection (DFS) pipeline, 3,058 microbiome samples (621) were age predicted with 5-fold 90-10% cross validation. Microbiome profiling is from open projects, available in ENA and SRA archives (project numbers: ERP019502, ERP009422, SRP002163, ERP004605, ERP002469, ERP008729 and ERP005534) the overall accuracy of this embodiment allows the present invention to predict the age of a person with a Mean Absolute Error (MAE) of 2.83 years.

Fig. 2A-2B show the observed and predicted ages for each age group in the DFS model (fig. 2A) and the XGB model (fig. 2B). MAE is the mean absolute error and N is the number of samples per age group. Thus, FIGS. 2A-2B provide a more detailed overview of the accuracy of DFS (FIG. 2A) and XGB (FIG. 2B) embodiments. The distribution of prediction errors among different age groups is not uniform, with the lowest prediction error among 30-60 years old.

Table 1 shows PFI importance, prevalence and average abundance in DFS and XGB models. This model may be referred to as model 1. Thus, table 1 provides information on the proposed set of microbial markers. The XGB and DFS importance columns show the decrease in R2 quality index (average of five equal points) after interchanging the corresponding features according to the PFI method. The prevalence bar shows the fraction of samples (cell number > 10/million) in which the signature was reliably detected, and the abundance bar shows the average relative number of microorganisms in the gut community in which the signature was detected.

Table 1:

in certain aspects, the microorganisms of table 1 are in a particular group and are used as a particular combination, where "group 95," as defined above, designates 95 particular members of the microorganism population. Accordingly, each of the 95 different types of microorganisms mentioned are present in the group 95.

Fig. 3 shows an embodiment for medical purposes. In particular, fig. 3 shows several operational embodiments of protocols that may be used by a medical professional. Classical metagenomic methods can be used to generate a taxonomic analysis of the gut microbiota of an individual, or in another microbiome of the individual. This profile was then analyzed using the present invention to determine whether the microbiota of the donor showed signs of accelerated aging. If the derived metagenomic clock of the present invention indicates that its microbiota is a typical significant senescent person, the software can pinpoint the specific changes that cause this effect. Subsequently, a health professional can evaluate this information (in conjunction with personal data) to provide lifestyle advice, select better treatment options for the patient, or plan a diet for the patient to support microorganisms that are "aging-retarding" (e.g., beneficial, associated with phenotypic youth) and inhibit microorganisms that are "premature aging" (e.g., harmful, or associated with phenotypic longevity) in the gut or otherwise.

FIG. 4 shows a comparison of the importance ratings of 74 microbial taxa assigned by the DFS and XGB age prediction models for all samples (0-90 years) and the age groups 15-30, 30-45, 45-60 and 60-90 years. Thus, fig. 4 illustrates the most important microorganisms for accurate age prediction in each age group, as shown by the PFI scores derived from the two age prediction models (XGB and DFS). After feature interchange of PFI scores²The higher the score of features, the more important the model performance. The 20 most important features for accurate prediction of age group samples from 15-30, 30-45, 45-60 and 60-90 years and all samples (0-90) were combined to form a list of 76 features. Their PFI score rating in each group was plotted on a heat map (black corresponding to the most important feature), rows (the importance vector for one feature within one model and age group) and columns were reordered according to manhattan distance to show similarity. The importance of features of the XGB and DFS models are highly similar, however, both rely on different features when predicting different age groups.

Table 2 shows the names associated with the numerical identifiers of fig. 4.

In certain aspects, the microorganisms of Table 2 are of a particular class and are used in particular combinations, where "class 76", as defined above, designates 76 particular members of that class of microorganisms. Accordingly, each of the 76 different types of microorganisms mentioned are present in the population 76.

Fig. 5 shows venn plots of the 100 most important microbial taxa for age prediction for both machine learning methods. Accordingly, fig. 5 illustrates the overlap between the 100 most important features in each age group as measured by PFI. Integers in parentheses indicate overlap of XGB derived features and integers in non-parentheses indicate overlap of DFS derived features. In both models, 40 to 41 features can be found in the 100 most significant feature list of all age groups, and no age group has more than 18 unique features in such a list.

Fig. 7 shows an embodiment of the operation applied to the insurance industry. Thus, FIG. 7 shows a simplified version of the invention as applied to an insurance embodiment. Based on the age prediction of the gut microbiome or other microbiome, a potential policy holder may be asked by the policy provider to modify the terms of the policy contract. Thereafter, the policy holder proceeds to a designated institution that performs microbiome screening and interpretation using the methods and/or markers specified herein. The test results are then stored on an immutable memory with an enabled smart contract, accessible to anyone possessing the sample donor ID. By sharing IDs, the donors allow policy providers to examine their gut microbiota (or other microbiome) levels or progression (e.g., over time) and corresponding age predictions. This information is then included by the policy provider in factors for calculating discounts on life insurance, or by third parties in evaluating available settlement fees. However, any type of insurance provider may use the present method of implementation. In addition, the implementation method can be modified for any industries, such as medical or contractual services and goods. The policy provider may alternatively be any type of contract administrator, modified based on the predicted biological age (i.e., the results of the biological clock model), and one or a series of results may be used to contract or modify any good or service.

Table 3 shows the performance of the gradient boost classifier. Thus, table 3 illustrates the predictive capabilities of the named tag lists provided herein. The gradient boosting classifier was trained on 75% of the available data and tested on 842 samples, using only the relative abundance of 95 (out of 1,673) microbial taxa. The quality indicator is calculated as set for the one-to-many algorithm. TP represents true increment (number of samples correctly assigned to the age group), FP represents false increment (number of samples incorrectly assigned to the group), FN represents false decrement (number of samples belonging to the group but assigned to another group), TN represents true decrement (number of samples correctly unassigned to the group). Accuracy equals TP/(TP + FP) (measure the fraction of all allocations correctly assigned to the group), failure equals FN/(FN + TP) (measure the fraction of test samples that should have been assigned to the group, but assigned to another group), accuracy equals (TP + TN)/(number of samples) (measure the number of correct allocations in all samples in the case of one-to-many). The P value is defined as the portion of the prediction vector arrangement that is better than observed. No such permutation combination provides better accuracy over 10'000 iterations.

Table 4 shows the random (equivalent) distribution of samples among the age groups. Thus, table 4 provides a frame of reference for table 3. These indices result from an equal ratio assignment for each age group sample. The P value is calculated as the number of permutations of the original prediction vector that provide better accuracy than the equal ratio assignment. Some random permutation combinations provide better classification accuracy, indicating that the gradient boosting model (table 3) performs significantly better than the equal ratio assignment.

TABLE 3

TABLE 4

Age group	15-30	31-60	61-90	Total number of
					Number of samples	378	265	199	842
TP	134	98	68
					FP	144	189	211
FN	244	167	131
					TN	320	390	432
Accurate	0.48	0.34	0.24
					Missing newspaper	0.65	0.63	0.66
Accurate and accurate	0.54	0.58	0.59
					P-value	<0.0001	<0.0001	<0.0001

FIGS. 8A-8X show ALE plots of 24 features in the DFS model. FIG. 8Y shows the calculated ALE values of FIGS. 8A-8X. Thus, FIGS. 8A-8X contain ALE plots of 24 features selected from the 100 most important features in the DFS model. The X-axis shows the relative abundance of a certain microorganism and the Y-axis shows the change in cumulative local effect (ALE) in years. ALE shows the average change predicted when adding the target feature. The dots represent the magnitude of the feature distribution (ALE is a magnitude of 1% to 95% calculated in 5% increments). The ALE is only calculated from samples whose target eigenvalue is non-zero. In certain aspects, the microorganisms shown in this paragraph and in FIGS. 8A-8X are in a particular group and as a particular combination, the "group 24" herein designates 24 particular members of the group. Accordingly, each of the 24 different types of microorganisms mentioned are present in the group 24.

Figure 12 shows calculated ALE values for another set of microorganisms. In certain aspects, the microorganisms of this paragraph are of a particular group and are used in particular combinations, and "group 19" herein is 19 particular members of that group. Accordingly, each of the 19 different types of microorganisms listed above are present in the group 19.

Figure 13 shows calculated ALE values for another set of microorganisms. In certain aspects, the microorganisms of this paragraph are of a particular group and are used in particular combinations, and "group 13" herein is 13 particular members of that group. Accordingly, each of the 13 different types of microorganisms listed above exists in the group 13.

Fig. 9A and 9B provide several indications of the use of ALE maps. Accordingly, fig. 9A and 9B include visual indications to explain the ALE map. FIG. 9A is an example of a potential senescence-delaying microorganism whose increase in abundance correlates with a predicted decrease in age. Fig. 9B is a potential early-senescence microorganism whose increase in abundance produced higher predictive values on average. In this particular example, the [ Eubacterium ] halii abundance increased by 0.8%, resulting in an average increase in predicted age of 0.57 years.

Fig. 10A and 10B show exemplary age distributions in a sample. Thus, fig. 10A and 10B illustrate the age distribution of data used for model training at the sample (fig. 10A) and individual (fig. 10B) level.

FIG. 11 shows a schematic diagram of an optimal configuration among neural networks tested for the application of DFS. Thus, FIG. 11 includes a schematic diagram of an optimal configuration for DFS applications among neural networks tested. There are 3 hidden layers, each with 512 neurons, set to a learning rate of 0.001 using the PReLU activation function. "x" is the input vector of features (i.e., relative abundance of microbial species).

In certain embodiments, kits are provided for the collection of microbiota material and subsequent metagenomic sequencing for the purpose of assessing the phenotypic age of a subject (e.g., a sample donor). Such kits may include instructions for employing a metagenomic sequencing protocol based on a human microbiome (e.g., gut microbiome). The kit can also include nucleic acids of the disclosed microbial marker sets. The kit may include guidance protocols that may vary in sample preparation, sequencing platforms and strategies, as well as quality control measures, normalization techniques, taxonomic assignments, and abundance calculation algorithms.

In certain embodiments, application of the methods described herein can be used to assess other chronological-dependent organism attributes, wherein the difference in calculated biological age as compared to chronological age is used to determine the presence of a disease state. For example, such methods may be used to estimate the risk of a disease or condition associated with aging, or to provide medical advice based on a predicted phenotypic age.

In certain embodiments, these methods can use a full set of disclosed microbial markers, as well as a subset thereof or a collection of higher order taxa consisting of the markers. For example, certain embodiments of the procedures may utilize amplicon-based sequencing techniques that, while not providing species-level resolution, may be used to assess community-level microbial community profiling.

In certain embodiments, the present invention can be used for medical purposes. Thus, the invention can include kits and tubing for using the disclosed set of labeled microorganisms to provide advice regarding daily work and diet as well as other factors, such as therapy or changes in daily work or diet. The present invention may assist a medical professional in selecting a drug, therapy or dietary intervention depending on the composition of, or resulting functional capacity of, a patient's microbiome (e.g., gut microbiome).

In certain embodiments, specialized databases, consulting software, extensions to existing patient identification systems, health guidelines, kits, and pipelines (and laboratories utilizing these) can derive age-based measures with the aid of the disclosed marker sets and methods described herein.

In certain embodiments, the microbial markers and methods may be used in a forensic setting to verify or determine the identity of a person. Forensic embodiments may include, but are not limited to, legal-capable certification systems, expert witness agencies, and laboratories that provide metagenomic-based age verification. All of these embodiments can utilize the above protocol to utilize information contained in one's microbial flora (e.g., gut) metagenome.

In certain embodiments, the microbial markers and methods can be used in a scientific context to gauge the effectiveness of novel therapies. For example, gut microbiome-derived age changes may be used to help identify the most effective drug candidates in compound screening studies with anti-aging properties.

In certain embodiments, a microbial clock may be generated in conjunction with microorganisms and methods to track and compare the rate of aging in different populations and/or different individuals. Such microorganisms can always be used in combination with biological pathway information, metabolic or strain analysis of gut metagenome or host genome information to generate new hypotheses in the fields of geriatrics, microbiology, immunology and general molecular biology.

All examples of the use of the disclosed microbial markers may vary in protocol details, accuracy, legal status, end purpose, etc., but are still based on the core concept of the chronological age and/or phenotypic age being linked to the indicators contained in the subject's microbiota, such as the intestinal microbiota. As such, all future practical applications of the present invention may be modified while still being limited to the markers and methods described herein. Thus, the list of possible operating embodiments is not limited to the embodiments described herein, but rather extends to those skilled in the art.

In certain embodiments, a method for predicting the phenotypic age of an individual based on the taxonomic profiling of the individual's microbiome is provided, as follows. (a) Isolating DNA or other nucleic acids from a microbiome sample of an individual, or otherwise obtaining nucleic acids; (b) estimating the relative abundance of a microbial taxa in the sample by WGS techniques to generate taxonomic features for the microbial flora, or otherwise obtain the estimated relative abundance; and (c) processing the classification features using a machine learning platform to predict a phenotypic age of the sample donor. Taxonomic profiling can include the microorganisms described herein, or their collections in different tables or different numbers. In certain aspects, step (c) uses derivative information of the microorganism, such as absolute abundance, genus level profiling, taxa-specific gene counts, ecogroup abundance, and the like. In certain aspects, the biological sample is processed to provide a taxonomic profile with resolution of strain, species, or genus. Alternatively, such taxonomic profiles are provided as data of a database. In certain aspects, the taxonomic profile is based on 16S variable region or whole metagenomic reads. In certain aspects, step (b) produces a taxonomic profile of relative or absolute abundance. In certain aspects, the taxonomic profile obtained in step (b) is used to compose a corresponding functional profile of the donor microbiome. In certain aspects, the sample in step (a) is a sample of the parenteral microbiome (e.g., genitourinary, skin, oral cavity), or the intestinal microbiome. In certain aspects, the sample donor is a human, and may also be a non-human animal.

In certain embodiments, the method may include comparing the metagenomic profile to a reference profile to find changes in microbiota that produce predictions of different ages, that may affect the predictions, and reporting the suggested changes to the sample donor.

In certain embodiments, a method may include compiling a metagenomic profile with personal data, such as images, medical history, biometrics, geographic location, and the like, to explore correlations between predicted age and the personal data.

In certain embodiments, the sample is a metagenomic gut, obtained by stool collection or biopsy.

In certain embodiments, the method may comprise using the obtained age prediction as a measure of phenotypic age. The method can also include reporting the predicted phenotypic age to the sample donor. The predicted phenotypic age may be compared to the chronological age of the subject, and the difference may be used to predict health.

In some embodiments, the method may also include identifying at least one microorganism or microorganism group whose abundance changes to affect the prediction of the sample donor and including this information.

In certain embodiments, the method can include comparing the data of the sample donor to a reference database, assessing the risk of having an age-related disease, and reporting information about the risk to the sample donor.

In certain embodiments, the method may comprise creating a dietary plan aimed at supporting or inhibiting a particular microorganism whose abundance change would affect the prediction, and providing the plan to the sample donor.

In certain embodiments, the method may comprise generating a recommended therapy for the individual based on comparing the individual's metagenomic characteristics and age prediction to a reference database, and reporting the therapy thereto.

In certain embodiments, the method may comprise evaluating multiple metagenomic profiles at once to find co-determinants of microbiome senescence and using this information to suggest changes in microbiota that may affect a population of people. Which may include modulating the predicted age. These methods may include creating a database of reference profiles and their corresponding predictions. The method may include information on the effect of at least one specific microorganism on age prediction, which allows for the creation of microbiota interventions that can alter the prediction, the development of such interventions, and the provision of microbiota interventions to a sample donor, subject or general public.

In certain embodiments, the method may include designing a consumer product based on the summarized information in the reference database and delivering the product to the consumer. In certain aspects, the methods can include creating a cosmetic product designed to support or inhibit a particular microorganism for which a change in abundance is predicted to affect, and providing the product to a sample donor, subject, or the public. In certain aspects, the methods can include creating a diet product designed to support or inhibit a particular microorganism for which changes in abundance are predicted to affect, and providing the product to a sample donor, subject, or the public. In certain aspects, the methods can include creating a medical product designed to support or inhibit a particular microorganism for which a change in abundance is predicted to affect, and providing the product to a sample donor, subject, or the public. In certain aspects, the methods can include creating a garment designed to support or inhibit a particular microorganism for which a change in abundance is predicted to affect, and delivering the product to a sample donor, subject, or the public.

In certain embodiments, the method may include creating software that compares the user's microbiological characteristics to a reference database and provides information about the effect of particular microorganisms on the user's age prediction and reports such information to the user. In certain aspects, the database is implemented as a blockchain storage to track the progress of the user's microbiological features and their derived age. In some aspects, personal data is shared only under a protocol that is guaranteed by its owner's encryption.

In some embodiments, the method can be used to develop insurance policies based on analyzing a reference database and providing them to customers. In some aspects, the rate of payment by the policy holder is influenced by age predictions derived from their microbiological profiling. In certain aspects, policy holders are expected to provide validated dynamics of their microbiological profiling, and corresponding age prediction.

The present invention further provides a method of predicting the age of a subject based on age-related microbiota, comprising (a) obtaining a biological sample or nucleic acid thereof from the subject; (b) determining the amount or percentage of one or more genes associated with an age-related microbiota marker, which amount or percentage varies with age, or obtaining data thereof from a database or from off-site analysis. (c) Comparing the number or percentage of one or more genes associated with an age-related microbiota marker that varies with age with an age-related reference population of the same microbiota; and (d) obtaining a value or range of values that predicts the age of the subject. Wherein the number or percentage of one or more genes associated with one or more age-related microbiota markers that vary with age are compared to the expression of the same gene from an age-related reference population, including any statistical, multiple regression, linear regression analysis, tabular or graphical method, for predicting the age of a subject based on the expression of genes associated with age-related microbiota markers that vary with age, thereby predicting the age (e.g., phenotypic age) of the subject.

In certain embodiments, the method may include developing a certain drug therapy based on the output predicted phenotypic age. In certain aspects, the method may include developing an anti-aging therapy based on the generated output to predict the phenotypic age. In certain aspects, the method may include developing a certain aging repair therapy based on the generated output to predict the phenotypic age.

In certain embodiments, a method of predicting a phenotypic age of a subject based on microbiota taxonomic profiling of the subject's microbiome can comprise: isolating a plurality of microbial nucleic acids from a subject microbiome sample, or otherwise obtaining nucleic acids; the plurality of microbial nucleic acids are analyzed to identify a microbial population of the microbiome based on the plurality of microbial nucleic acids, or to otherwise receive data regarding the microbial population. Generating a taxonomic profile of the subject microbiome based on the number of each microorganism or otherwise receiving the taxonomic profile data; processing taxonomic profiling of microbiota using a computer configured with a machine learning platform (e.g., the machine learning platform includes one or more deep neural networks) to predict chronological age by calculating phenotypic age of the subject; generating a report with the predicted chronological age and/or the calculated phenotypic age of the subject; and providing the report to the subject.

The predicted chronological age, i.e. the identified phenotypic age, is also referred to as the biological age. The predicted chronological age may be compared to the actual chronological age, and the difference may be used to identify the patient's health condition. The lower the phenotypic age, compared to the chronological age, the more likely the subject is healthy. Larger differences between phenotypic age and chronological age can be used to identify disease states.

In certain embodiments, the processing of microbiome taxonomic profiling results in defining one or more of: absolute number of microorganisms; absolute number of each microorganism; taxonomic profiling of genus fractions of absolute numbers of individual genera of microorganisms; (ii) a species-level taxonomic profile of the absolute number of individual species of the microorganism; a strain level taxonomic profile of the absolute number of each strain of the microorganism; a taxon-specific basis for the absolute number of microorganisms of the taxon; absolute number of ecological groups. The relative number of total microorganisms in the microbiome; the relative number of each microorganism; the relative number of each microorganism genus; the relative number of each microbial species; the relative number of each microbial species; the relative number of each microbial species; the relative number of each microbial species; relative number of taxa microorganisms taxa-specific basis factors; the relative number of ecological groups; or a combination thereof.

In certain embodiments, a method may include generating one or more of: a species-level taxonomic profile of the subject microbiome based on the number of each microorganism; a species-level taxonomic profile of the subject microbiome based on the number of each microorganism; or taxonomic profiling of genus classes of the subject microbiome based on the number of each microorganism.

In certain embodiments, each taxonomic profile is based on at least one of the 16S variable regions, or on the entire metagenomic read.

In certain embodiments, a method may comprise: accessing a database of a plurality of reference microbiologic profiles, the database being associated with chronological ages and/or phenotypic ages of a plurality of reference subjects; comparing the taxonomic profile of the microbiome of the subject to a plurality of reference taxonomic profiles of microorganisms; and grouping one or more microorganisms associated with the predicted phenotypic age range.

In certain embodiments, a method may comprise: identifying an altered taxonomic profile to reduce the predicted phenotypic age of the subject to a predicted phenotypic age range that is lower than the predicted phenotypic age of the subject; and including the altered taxonomic profile in a report provided to the subject.

In certain embodiments, a method may comprise: identifying a process for obtaining an altered taxonomic profile of the subject to obtain a younger predicted phenotypic age range in the subject; administering a treatment method to the subject; and obtaining a younger predicted phenotypic age range in the subject.

In certain embodiments, a plurality of reference microbiologic profiles associated with chronological and/or phenotypic ages of a plurality of reference subjects are associated with personal data of the reference subjects. Accordingly, the method may comprise: obtaining personal data of a subject; correlating the personal data of the subject with the generated taxonomic profile of the subject and/or the predicted phenotypic age of the subject; in view of a plurality of reference microbiologic profiles associated with chronological and/or phenotypic ages of a plurality of reference subjects, a correlation of the predicted phenotypic ages of the subjects is identified in association with personal data of the reference subjects.

In certain embodiments, a method may comprise: identifying at least one microorganism whose change in quantity provides an altered taxonomic profile to reduce the subject's predicted phenotypic age to a predicted phenotypic age range below the subject's predicted phenotypic age; altering the number of at least one microorganism identified in the subject's microbiome; and obtaining a range of predicted phenotypic ages of younger in the subject.

In certain embodiments, a method may comprise: accessing a database having a plurality of reference microbiologic profiles associated with a plurality of diseases and/or disorders of a plurality of reference subjects; comparing the microbiome taxonomy profile of the subject to a plurality of reference microbiome taxonomy profiles; identifying a subject at risk of developing a plurality of diseases and/or one or more conditions thereof; and reporting the risk of the subject developing the plurality of diseases and/or one or more disorders thereof.

In certain embodiments, a method may comprise: identifying a dietary plan for obtaining an altered taxonomic profile of the subject to obtain a lighter predicted phenotypic age range in the subject; and incorporating the diet plan into the report. In certain aspects, a method may comprise: administering a dietary plan to the subject; and obtaining a lighter predicted phenotypic age range in the subject.

In an embodiment, a method may comprise: identifying a therapeutic treatment composition having one or more microorganisms for obtaining an altered taxonomic profile of a subject to obtain a lighter predicted phenotypic age range in the subject; and determining the therapeutic composition in the report. In certain aspects, a method may comprise: administering a therapeutic treatment composition to the subject; and a lighter predicted phenotypic age range in the subject.

In some embodiments, a method of generating a reference database may be provided, which may be used in any method, which may include: isolating a plurality of microbial nucleic acids from each of a plurality of reference samples, each reference sample from a microbiome of a reference subject; analyzing the plurality of microbial nucleic acids to identify the number of microflora for each reference subject based on the plurality of microbial nucleic acids. Generating a taxonomic profile of the microbiome for each reference subject based on the number of each microorganism; processing a taxonomic profile of the microbiome of each reference subject using a computer configured with a machine learning platform to predict the phenotypic age of each reference subject; and generating a reference database using the predicted phenotypic age associated with the microbiome taxonomic profile of each reference subject.

In certain embodiments, a method may comprise: analyzing a plurality of reference microbiologic taxonomic profiles associated with age and/or phenotypic age of a plurality of reference subjects; and identifying one or more common microorganisms in the plurality of reference taxonomic profiles associated with a higher predicted phenotypic age for the plurality of reference subjects, and/or identifying one or more common microorganisms in the plurality of reference taxonomic profiles associated with a lower predicted phenotypic age for the plurality of reference subjects.

In certain embodiments, a method may comprise: identifying at least one altered taxonomic profile to reduce the predicted phenotypic age of the reference subject to a predicted phenotypic age range that is lower than the reference subject's phenotypic age; incorporating the at least one altered taxonomic profile into a report; and providing a report to a plurality of reference subjects.

In some embodiments, a method may comprise: identifying at least one specific microorganism that modulates the phenotypic age of a reference subject; creating a regulatory composition for regulating at least one specific microorganism identified in a microbiome; and providing the conditioning composition to at least the subject.

In certain embodiments, the method can include generating a reference database having a predicted phenotypic age associated with a microbiome taxonomic profile of a plurality of reference subjects.

In some embodiments, a method may comprise: analyzing a plurality of microbiology profiles from a plurality of subjects in view of a predicted phenotypic age associated with the microbiome taxonomic profiles of the plurality of subjects; identifying at least one common microorganism affecting the predicted phenotypic age; identifying at least one microbiota alteration to alter at least one first microbiologic profile of at least one first subject to obtain a lower predicted phenotypic age; and optionally providing the identified change in the at least one microbiota to at least one first subject in the form of a first report.

In certain embodiments, the method can include generating a database having a plurality of taxonomic profiles of microorganisms associated with a predicted phenotypic age associated with a taxonomic profile of a microbiome of a plurality of subjects.

In certain embodiments, a method may comprise: creating a microbiota intervention to obtain at least one microbiota alteration to alter a first taxonomic profile of microorganisms of at least a first subject to obtain a lower predicted phenotypic age; and providing the microbial intervention to at least a first subject. In certain aspects, a method may comprise: designing a consumable for microbiota intervention; and providing the consumable product to at least one consumer. In certain aspects, a method may comprise: designing a dietary product for cluster intervention; and providing the dietary product to at least one consumer. In certain aspects, a method may comprise: designing a medical product for microbiota intervention; and providing the medical product to at least one consumer. In certain aspects, a method may comprise: designing a garment for microbiota intervention; and providing the laundry to at least one consumer. In these methods, the subject can utilize the provided article to reduce its expected phenotypic age.

In certain embodiments, a method may comprise: generating a computer program product stored on a tangible, non-transitory storage device of a computer, the computer program product, when executed, causing the computer to: accessing a database according to embodiments described herein; and comparing the taxonomic profile of the subject to the database; providing information about at least one specific microorganism that modulates a predicted phenotypic age associated with the subject; generating a report containing the provided information; the report is ultimately provided to the subject. In certain aspects, the database is configured as a blockchain storage system that is capable of tracking changes in taxonomic profiles of the subject and/or subjects, as well as tracking changes in the predicted phenotypic age of the subject or subjects.

In some embodiments, a method of generating an insurance policy may include: accessing an insurance profile database, the insurance profile database having a plurality of taxonomic profiles of microorganisms; the taxonomic profile of microorganisms is correlated with a predicted phenotypic age that correlates with the taxonomic profile of each microbiome of the plurality of subjects. Obtaining information of insurance clients; comparing the information of the insurance client with the insurance file database; generating an insurance policy based on the comparison of the insurance client's information with the insurance file database; and provides the insurance policy to the insurance client. In certain aspects, a method may comprise: obtaining an insurance client predicted phenotypic age resulting from microbiologic profiling of the insurance client; and obtaining a predicted phenotypic age of the insured customer. Determining a rate for the policy based on the predicted phenotypic age of the insured customer; and include the rate in the policy. In certain aspects, an insurance client predicted phenotypic age resulting from microbiologic profiling of an insurance client is obtained by: isolating microbial nucleic acids of a plurality of microorganisms from a sample of a microbiome of an insurance client; analyzing the plurality of microbial nucleic acids to identify a microbial population of the microbial community based on the plurality of microbial nucleic acids; generating a taxonomic profile of the insurance client microbial community based on the number of each microorganism; processing taxonomic profiles of microbiome using a computer configured with a machine learning platform to predict phenotypic age of the insured customer; generating a report containing the age of the insurance client's expected phenotype; and provides the report to the insurance provider managing the insurance policy. In certain aspects, the microorganism is several of those described herein.

In some embodiments, the insurance profile database is obtained in the following manner. Analyzing a plurality of taxonomic profiles of microorganisms from a plurality of subjects in view of a predicted phenotypic age associated with the taxonomic profiles of the microbiome of the plurality of subjects; identifying at least one common microorganism that affects the predicted phenotypic age; identifying at least one microbiota alteration to alter at least one first microbiologic profile of at least one first subject to obtain a lighter predicted phenotypic age; and storing the identified at least one microbiome change in an insurance profile database.

In certain embodiments, a method of deriving a microbial biomarker of aging may comprise: creating a machine learning model trained to predict host age using taxonomic profiling of microbiological samples; and selecting a subset of microbial features based on the cumulative local effect (ALE) function behavior (monotonicity and range) of the microbes.

In certain embodiments, any of the methods herein can be performed using a biomarker for a microorganism. The method may comprise using a biomarker for a particular group as defined herein, such that the group comprises all members of the defined group. In certain aspects, biomarkers and microorganisms include groups thereof, which are any group or group described in this or a co-pending application.

Certain embodiments of the present invention allow a person (e.g., a subject) to track the status of their gut microbiome through a web-based application. Here, the subject can provide a biological sample or biological data (e.g., nucleic acids, proteins, chemicals, etc.) obtained from analysis of a biological sample. Thereafter, the web-based application can provide the reports, data, and analysis described herein to the subject, such as providing information to the subject's computer (e.g., laptop, tablet, smartphone, handheld device). This information may be the same information as determined in the report and may include: a predicted biological age compared to the temporal age of the subject based on one or more individual samples from the subject; an interactive chart for selecting a defined age range; information of a defined age range; quality or statistics of predicted biological age (e.g. MAE, RMSE R)²Pearson's R, etc.); determining the number or type of biological ages); identifying the number and/or type of microorganisms used to identify a predicted biological age; minimum diversity coverage of the model; maximum diversity coverage of the model; comparison of any data or results between different samples; the predicted biological age is compared with other data or results over time in a plurality of samples. Information having a microbial appearance similar to that of the gut microbiome, to keep the identity of the individual as a microbial appearance secret; altering the nutritional recommendations for predicting the age of the organism; administering a probiotic to alter the identity of the predicted biological age; not consuming the probiotic to alter the identity of the predicted biological age; or other information.

The process of the web-based application may also include information regarding monotonically increasing and decreasing characteristics. This may include aggregating monotone increment and decrement features into other statistics to further reduce dimensionality. For example, all monotonically decreasing or increasing features may be replaced by their weighted average abundance, with weights representing their correlation with age, ALE amplitude, ALE rate of change, or other values. This results in the first 41 features of cluster 41 (or other number of other clusters) being collapsed into 2 dimensions, which can be rendered on a web interface. The user may view the map or provide it to a medical professional to determine the overall condition of his intestinal plexus. Thereafter, the application interface or professional may provide nutritional recommendations or medication to decrease the number of monotonically increasing (age-related) microorganisms or increase the number of monotonically decreasing (young-related) microorganisms.

In certain embodiments, the biological sample from the subject is a stool sample containing a microbiome. Thus, the method may comprise the subject ordering at least one fecal sample collection container and placing the fecal sample in the collection container. The fecal collection container can thereafter be sent to analyze nucleic acids of microorganisms from a microbiome sample from the subject. The web-based application entity may process the stool sample and may also send the stool sample to a dedicated stool sample facility. At some point in the process, the nucleic acids of the microbial flora need to be analyzed, regardless of the entity that is processing the fecal sample. The method steps of analyzing include generating a taxonomic profile of the microbiome, processing the taxonomic profile using a machine learning platform to obtain a predicted biological age of the subject, and generating and providing a report, such as displaying the report on a computing device of the subject.

FIG. 22 illustrates an embodiment of an application capable of performing the methods described herein. This example includes an application that can receive or provide a known number of individual microorganisms in a microbiome, or receive or provide a taxonomic profile of a known subject microbiome. In any case, the taxonomic profile is processed in order to obtain or predict a biological age (e.g., phenotypic age) of the subject. Finally, the application provides the displayed information to the subject in a report. The report may include a chart showing the scores of the patent target and other patent targets (e.g., your score). The chart may be interactive and allow the subject to adjust an estimated parameter, such as KDE, as a percentage defined in the population. The report may also include a chart or table of results relating to the biological clock model. The report may also include the identity of certain microorganisms or biomarkers of the microorganisms. This information can be presented so that the subject is informed of the degradable microorganisms or biomarkers that can lead to a reduction in the predicted biological age, and selection of a microorganism or biomarker can provide information on how to reduce gut microbes. Presentation of the information may inform the subject of increasable microorganisms or biomarkers that may result in a decrease in the predicted biological age, and selection of a microorganism or biomarker may provide information on how to increase gut microbiology. In the example of FIG. 22, the biomarker 6 is selected (e.g., by clicking or hovering over a selection icon), wherein increasing the information that the biomarker 6 is present includes: adjusting the diet in a specific manner; avoiding exposure to certain types of radiation; avoiding exposure to specific types of compounds; exercise a particular skill (e.g., physical activity, running, weight lifting, yoga, stretching, etc.).

In certain embodiments, a subscription-based service may be provided to a subject. The subscription may allow processing of multiple samples and identifying multiple biological ages from sequences of the multiple samples. This may allow tracking of the final score for each sample. To determine a regimen that may be able to reduce the age of a higher predicted organism, the trend of the sample and its changes may be analyzed. Then, the application can provide a prompt for intervention to the end user; these interventions may shift the outcome of the biological age to a lower predicted age, which may improve the health of the subject.

In a certain embodiment, the information of all determined biomarkers is compressed to 2 dimensions by the following procedure:

here "X" is the normalized (e.g., replacing abundance by population percentile in the 0-100% range) monotonically increasing feature abundance vector, while "w" is the weight vector (e.g., ALE plot numerical derivative, exact ALE value, or other statistic), and "Y" is the normalized abundance vector for the serum recessive feature. These two new values may be computed for a selected population (e.g., healthy individuals of the same age) to provide a Kernel Density Estimation (KDE) graph to illustrate the score distribution and to give the end user a frame of reference. The KDE field may be tailored to encompass a particular portion (e.g., 90% or 50%) of the selected population.

In certain embodiments, a method of creating a microbiomic aging clock for a subject, the method comprising: (a) receiving a microbial abundance profile of a subject microbiome; (b) creating an input vector according to the abundance characteristics of the microorganisms; (c) inputting the input vector to a machine learning platform; (d) generating, by a machine learning platform, a predicted microbiome senescence clock for a subject microbiome based on an input vector, wherein the microbiome senescence clock is specific to the microbiome; and (e) composing a report comprising a microbiomic aging clock, the report determining the predicted biological age of the subject. In certain aspects, the method may comprise, after a specified period of time: performing steps (a), (b), (c), (d), and (e) in a second iteration; and comparing the initial report with the report of the second iteration; and determining the predicted change in the age of the living being over a specified period of time.

In some embodiments, the method may comprise: creating at least a second microbiome senescence clock by repeating any one or more of steps (a), (b), (c), and (or) (d), wherein the second microbiome senescence clock is based on a second microbial abundance characteristic from the subject's microbiome, the subject's different microbiome, or the second subject's microbiome. And optionally preparing a report including a second microbiome aging clock; the report identifies a second predicted biological age of the subject, a different tissue or organ of the subject, or a tissue or organ of a second subject. In certain aspects, the method can include combining a microbiomic aging clock with a second microbiomic aging clock to create a synthetic microbiomic aging clock, wherein the synthetic microbiomic aging clock provides a synthetic biological age of the subject; alternatively, a report comprising a synthetic microbiome aging clock identifying the subject's synthetic biological age can be prepared. In certain aspects, the method may comprise: comparing the predicted biological age of the tissue or organ to the actual age of the subject; comparing the second predicted biological age of the tissue or organ to the actual age of the subject; comparing the synthetic biological age of the tissue or organ to the actual age of the subject; wherein the method further comprises: a report is prepared containing the comparison and the actual age difference from the subject.

In certain embodiments, the report includes one or more of: a treatment regimen based on the predicted biological age in view of the actual age of the subject; a dietary regime based on the predicted biological age in view of the actual age of the subject; questionnaires regarding lifestyle habits; prognosis of life expectancy with and/or without treatment regimen; prognosis of life expectancy with and (without) dietary regimens; prognosis of the likelihood of patient survival in a treatment regimen; or the likelihood of survival of the patient during the dietary regimen. In certain aspects, the method comprises: comparing the initial report with the report of the second iteration; determining a change in the predicted biological age over a defined period of time; and determining: whether the treatment regimen changes the predicted biological age, and if the treatment regimen changes the predicted biological age, determining whether: continuing the treatment regimen, changing the treatment regimen, or stopping the treatment regimen, or if the treatment regimen does not change the predicted biological age, determining whether: continue the treatment regimen, change the treatment regimen, or stop the treatment regimen.

In certain embodiments, the report may include a treatment regimen of the biological age predicted from the actual age of the subject, or a dietary regimen of the biological age predicted from the actual age of the subject.

In certain embodiments, the method may include performing one or more of: a subject-actuarial assessment based on the predicted biological age; a risk assessment based on the predicted biological age; or insurance assessment based on predicted biological age.

In certain embodiments, a computer program product comprises a tangible, non-transitory computer readable medium having computer readable program code stored thereon, the code executable by a processor to perform a method for a biological aging clock of a patient, wherein the method comprises the computer performing the steps of one of the methods described herein.

Examples

All data used in the DFS operation examples are from published studies (project IDs in SRA and ENA: ERP019502, ERP009422, SRP002163, ERP004605, ERP002469, ERP008729, ERP 005534).

All of these studies used similar sample preparation techniques. Stool samples (about 50 grams) were taken from healthy donors and if DNA isolation could not be performed immediately, the samples were stored at 4 ℃ or frozen in an environment at-80 ℃ to prevent bacterial growth.

DNA can be isolated using various kits and methods in preparation for WGS sequencing on the Illumina Genome Analyzer II or Illumina HiSeq 2000 platform (Wesolowska-Andersen et al, 2014). Reading processing and taxonomic profiling can be implemented. The downloaded sequencing file is not down sampled. Quality control includes aptamer removal, mass trimming and mass filtering, and human reading removal. In addition, the entropy of the 4-mer distribution was also evaluated to verify that none of the samples were over-diluted. The FM index against previously released bacterial and archaeal genomes was plotted against reads in the centrifuge. The relative abundance table for each sample was filtered to contain only reliably detected microorganisms (relative abundance >1 e-5).

Several machine learning methods were used for age prediction in this study. The goal of the machine learning problem is classical supervised regression. The model was trained with bacterial taxa characteristics as input variables.

The data used to train the machine learning algorithm was modified to contain only reliably detected features (>1e-5 abundance). After removing the age values that were absent or missing, the data was scaled to 3'058 records from 621. Each record contained 1'673 bacterial signatures. The target variables include an age in the range of 7 to 90 years. The age distribution is depicted in fig. 12.

The training of the model employs quintuple cross validation and a grid or random search strategy to compensate for the overfitting. Since the size of the data set is rather modest, no external test set is introduced. Instead, the given index is a prediction for each aliquot of 5 test sets, resulting in a fully predicted input data set. For a fair comparison, all models were evaluated using the same polyline segmentation method.

All methods showed relatively good correlation of predicted age with actual age, however both gradient lifters and DFS performed better than the other models. The optimal configuration of the DFS model has 3 layers with 512 neurons per layer, a prellu activation function and a learning rate of 0.001. The maximum depth of the tree of the optimal gradient lifting configuration is 6, the learning rate is 0.1, the maximum delta step length is 2, the second-level regularization term is 0, and the first-level regularization term is 0.5. Table 5 lists CV results for the XGB and DFS models.

TABLE 5

Model (model)	MAE	RMSE	R2	Pearson
					XGB	3.81	5.63	0.90	0.95
DFS	2.83	5.00	0.92	0.96

Method for constructing microbiology aging clock

As described in the following steps, and as shown in fig. 14, a method 1400 of constructing a microbiomic aging clock is provided. The method includes obtaining nucleic acid information contained in the intestines of a group of subjects (element 1402). This can be done by obtaining nucleic acids from a biological sample from each subject and detecting them to obtain information, or can be obtained elsewhere and provide nucleic acid information for the method. All subjects should have a definite Chronological Age (CA). Age distribution in homogeneous groups affects the final clock bias, so a uniform distribution is recommended. The cohort should include subjects within the age range of expected use, for example a clock used for training in the younger cohort would produce a reliable prediction for younger subjects but not for older subjects. In practice, published research data is used, with ENA identifiers being: ERP002061, ERP008729, SRP002163, ERP004605, ERP003612, ERP002469, ERP019502, SRP008047, ERP009422, ERP005534, PRJEB2054, PRJNA375935 and PRJNA 289586. The data included 1,165 people in the age range of 20-90 years, but in the accompanying study metadata obtained from the ENA or supplemental information, health status was not specified. Figure 24 shows the age distribution of 1,165 people in model 2.

The method 1400 may include filtering the obtained nucleic acid information (element 1404). In certain aspects, all WGS reads are truncated according to the quality score of the terminal base (cut-off point: 50Phred score), and all reads with an average quality less than 20Phred score are completely deleted. Furthermore, all reads derived from the human genome were deleted using the Bowtie2 alignment tool.

The method 1400 may include inferring a microbial abundance profile from the nucleic acid information (element 1406). In certain aspects, the centrifuge classifier can be used for metagenomic sequences with bacterial, archaeal indices. The centrifuge single output file is summarized to a sparse table containing the relative abundance profiles of the microbial species for all subjects in the cohort. The relative abundance profiles may be obtained by other software products (such as Metapthlan 2) or by normalizing the absolute abundance profiles.

The method 1400 may include filtering and normalizing the raw abundance distribution (step 1408). Each microbial taxonomic group in the relative abundance table is considered herein to be a mathematical feature. To reduce the chance of over-fitting of subsequent models, the approach involves reducing dimensionality, leaving only the most relevant features. There are a variety of feature selection methods available for this task. In certain aspects, this can include assigning zero values to all abundances <1e-5 using a direct and simple method, which is demonstrated to remove unreliable detected microorganisms. Thereafter, the protocol removed all features present in <0.0013 samples, resulting in no one sample losing > 5% of the population, then all values were divided by the sum of the abundances in their respective samples.

Method 1400 may include defining a cross-validation dataset (element 1410). To reduce the effects that may result from data bias for any particular study, the data is divided into 10 equal portions, and the studies are stratified by the source of the data. During the training phase, all the aliquots remain fixed. The separated data may be used as a deterministic data set, and the model may be validated by a modeling process. Different data sets may be used for verification.

The method 1400 may include training a neural network model (element 1412). Some neural network architectures may be selected by professionals and trained with various parameter configurations. Neural network training was performed with the public library of Python3, Keras and tensrflow. All possible configurations were tried using a grid search method. Since cross-validation is used at this stage, each parameter and architecture configuration is used to generate as many models as possible in equal parts. In this training, each configuration produced 10 models trained on different pieces of raw data. Training lasts 2,500 durations.

The method 1400 may include evaluating model performance (block 1414); in this case, 10 models are generated for each parameter configuration. Each parameter is used to predict a respective test set as defined in the cross-validation score definition phase. The overall quality of the configuration is evaluated by using the Mean Absolute Error (MAE) of all the aliquots; other indicators, such as R, may also be used²Pearson's R, mean square error or median absolute error. The best configuration obtained in cross-validation is: 3x1024 nodes, ReLU activation function, Adam optimizer, discard per layer with a probability of 0.5 (Dropout), and a learning rate of 0.001. This is from model 2.

The method 1400 may include generating a set of models having models with low errors (block 1416). The models that exhibit the smallest error are combined into a collection, which is then used on the previously untouched validation data set, after which the predictions of the collection model are averaged by taking their average prediction values. Other aggregation methods may also be effective; for example, the median or outlier prediction is removed and summed, or the midpoint of the prediction range is selected. The final model error is reported as the best configured MAE-5.91 years. This is from model 2.

Assessing the effect of probiotics on host condition

The biological clock model described herein can be used to assess the effect of probiotics on the condition of the host, which can include the assumption that administration of probiotics only alters the abundance of probiotics. Thus, using this model it can be easily examined how a particular microorganism affects the intestinal age. The geriatric potential of 16 commercial probiotics was assessed by model 2, and the microorganisms were as follows: bacillus coemulans, Bifidobacterium animalis, Bifidobacterium bifidum, Bifidobacterium breve, Bifidobacterium longum, Enterobacter failure, Lactobacillus acidophilus, Lactobacillus casei, Lactobacillus gasseri, Lactobacillus helveticus, Lactobacillus plantarum, Lactobacillus reuteri, Lactobacillus rhodo, Lactobacillus sanctual, Lactobacillus fruterium, and Streptococcus thermophilus. However, the effect of any other microorganism whose abundance is among the model inputs can be examined using the method. In certain aspects, the microorganisms of the present paragraph are of a particular group and are used in a particular combination, where "group 16" designates 16 particular members of the group. Accordingly, each of the 16 different types of microorganisms mentioned are present in the group 16.

To illustrate how these probiotics influence age prediction at the individual level, the protocol increased their abundance by 1% in seven relative abundance profiles and calculated changes in predicted age. Other numerical differentiation methods may also be applied. Using this model, the efficiency of implantation of probiotics cannot be estimated, and therefore the absolute value of the predicted age change is not as important as the signs of change. However, signs of age change can be used to predict the ability of probiotics to elicit geriatric protection. Fig. 15 shows the simulated effect of increasing 16 probiotics in microbiota profiling, where dark grey is predicted age increase, light grey is predicted age decrease and white is predicted age no change.

As can be seen from the example of fig. 15, by using the biological clock model, different people can have a positive and negative impact on the predicted age of the same bacterium. To understand how certain bacteria affect gut age at the population level, the use of cumulative local effect (ALE) is recommended. Methods of calculating ALE are provided herein.

The data shows that some of the microorganisms used in the model were determined to be monotonically decreasing in the population >50 years of age. Increasing the age prediction that some of them have almost linear effects, while others only have an effect at the high end of their abundance distribution, some may have a non-monotonic effect on the population. The results in the tables of fig. 16A and 16B are obtained by processing the disclosed ENA entries and provide proof of concept.

The data in fig. 16B shows the effect of 6 microorganisms on sampling of the population over 50 years of age. The change in ALE between the two points on the X-axis indicates the average change in predicted age, together with the shift in abundance of a microorganism. In each particular case, only people with a non-zero microbial abundance are considered.

The data in fig. 16A shows the predicted mean change while shifting the abundance of microorganisms in a sub-sample of subjects over 50 years of age. In each particular case, only people with a non-zero microbial abundance are considered.

Predicting disease and risk of onset

The biological clock model can also be used to assess the intestinal age for incidence prediction. This can also be used to predict the risk of obesity. For example, by using a biological clock model and microbiology information, a Body Mass Index (BMI) can be predicted.

In certain embodiments, a biological clock model may be used to predict BMI values. The intestinal age predicted by the biological clock model can be used together with microbiology information for predicting BMI value of a person as a substitute obesity risk index. Using information on the abundance of the 314 prevalent microbial species (as defined in the centrifuge index) and the predicted gut age, the model successfully estimated the BMI value of the host over the baseline median distribution. Specific microorganisms are:

the sample used to train and validate the BMI predictor contained 968 microbiome profiles with known BMI values and age prediction error <10 years. Fig. 23 provides an example of 30 samples showing sample ID, actual BMI value, predicted phenotypic age, and actual age. These data were also used for training of the aging clock, whose predicted age was identified during the cross-validation phase. These profiles are taken from samples stored in the ENA. A particular version of the centrifuge index is "from bacteria, archaea" last updated in 2018, 4 and 15.

MAE produced by median BMI allocation was 4.129kg/m²While the MAE of the biological clock model is 3.642 + -0.208 kg/m2, which is estimated in five aliquots. The BMI value estimator is implemented as a gradient boost regressor, created with the XGBoost 0.90Python 3.7.3 library. The exact parameters used during training are: { alpha:0, eta 0.1, eval _ metric: mae, gamma 0.0, lambda 0.5, max _ delta _ step:0.0max _ depth:6, tree _ method: gpu _ hist }, which are provided as examples. However, variations thereof may also be used.

Fig. 17 shows a plot of the Kernel Density Estimate (KDE) of BMI predictions obtained with the XGBoost model, which is applicable to 968 relative abundance profiles with accompanying information on gut age. The line represents the common least squares regression of the observed values with the equation (BMI predicted value) × (actual BMI value) + 20.16.

In certain embodiments, the biological clock can be used in a method for predicting the risk of developing type I diabetes (T1D), which can use model 2. In validating an intestinal age predictor containing 1,606 features, the protocol was modified to include the data obtained from the T1D patient in the validation set. According to the biological clock model trained with T1D data, the predicted age of T1D patients is more likely to be higher than the actual age, with an average error of +14.75 years. By comparison, the data for healthy hosts in this study (PRJNA289586) had an average prediction error of + 0.41. While the MAE of the T1D host was 18.02 years old, the MAE of the healthy host was 8.40 years old. Thus, the error profile of a diabetic patient is biased towards a high positive number, which can be used as an indicator of his health status and thus as a biomarker for diabetes.

Figure 18 shows the performance of model 2 in a separate set of 436 samples from a public study (ERP000108, PRJNA375935, PRJNA289586) stored in an ENA server, the latter containing 34 samples from donors with type 1 diabetes. Median age assignment provided a baseline for Mean Absolute Error (MAE) of 9.27 years, while the model showed MAE of 5.91 years (R2 score ═ 0.81). At the same time, the model failed to accurately predict the age of the diabetic (R2 score-0.73, MAE 18.02, baseline 13.65) and the age assigned to the young diabetic donor was higher than its actual age. This indicates that the model qualifies as an accurate aging clock for healthy people and reacts to disease as a condition for age influence. FIG. 18 includes a validation set (N) for age prediction_Total＝323，N_T1D34), the model contains 1,606 features. The dotted line represents the data of the diabeticThe OLS regression of the material, the solid line is the OLS regression of the healthy person data. The predicted age of a diabetic patient is usually higher than that of a healthy person.

Biological aging clock model

The following example provides a method for reducing dimensionality of high throughput biological data and ranks features according to their magnitude of impact on the aging process. The method may include two parts, as shown in fig. 19: (1) training a reliable aging clock model, and (2) applying cumulative local effects (ALEs) to the model to measure its response to changes in the feature values. The initial aging clock can be trained on any biologically relevant data, the only limitation being that it is a finite predetermined dimension (e.g., a vector of relative abundance of microorganisms). The model may represent essentially any method of machine learning, including the Deep Neural Network (DNN) used in the described embodiments.

In this example, the DNN model was trained to predict a person's age based on their intestinal plexus relative abundance profiles. The DNN model consists of 3 hidden layers, each layer having 1,024 nodes, trained using the prilu activation function, with a drop (Dropout) rate of 0.5 and a learning rate of 0.001. Only 20-90 year old population samples were used, and only one sample per donor was used, giving a total of 1,165 donors.

During validation of the independent data set, the mean absolute error for this model was 5.91 years, with baseline (median age assignment) yielding an error of 9.27 years. The predicted age of type 1 diabetic patients is higher than the chronological age, resulting in an error of 18.02 years with a baseline of 13.65 years. Thus, the model accurately predicts the age of a healthy person and responds by overestimating the age of the patient.

Cumulative local effect Analysis (ALE) was then performed on the model to see how the model responds to slight changes in the abundance of a particular microorganism. If the model predicts an age by average reduction when modeling a microorganism to be more abundant, the microorganism is assigned a monotonically decreasing state. The model is monotonically increasing if it yields a higher predicted age with increasing abundance of microorganisms. For example, ALE equation 1 may be used to calculate the Local Effect (LE) for each quartile. To generate ALE, LEs will be added according to ALE equation 2.

Furthermore, the ALE is centered, so the sum of all ALEs of a feature is zero.

In this embodiment, the biomarkers of aging are selected according to the amplitude of the ALE, for example by changing some characteristic value to achieve a predicted maximum transition. Only features with amplitudes above 0.1 years of age were selected, resulting in a list of microorganisms of the population 41 as defined herein. However, other microbial populations as defined herein may also be used.

To estimate the effect of a certain feature on the model performance, all samples are separated from a single set into equal bins (Bin) according to a selected feature distribution (in this case, step size of 25%). The predicted age of each bin is then calculated at its left and right boundaries, and the predicted changes between bins are averaged to obtain the Local Effect (LE). The local effects are further summed up into a cumulative local effect (ALE) to continuously manifest the effect of a feature. In the last step, all cumulative local effects are shifted to make the average cumulative local effect zero. See fig. 19.

Most features typically show a monotonic increase or decrease when subjected to cumulative local effect (ALE) analysis. The feature showing that the predicted age increases with increasing value (increasing ALE plots) is called monotonic increase. On the other hand, the feature that the predicted age decreases with an increase in the numerical value (decreasing ALE map) is called monotonic decrease. Some features do not produce monotonic ALE maps, but most are monotonic. Fig. 20 shows ALE plots of left-most Bin (Bin) features with maxima and minima located at the feature distribution, i.e. monotonically decreasing and monotonically increasing features. Thus, fig. 20 provides examples of monotonically increasing and monotonically decreasing behavior of the aging clock of the gut microbiota, as well as several non-monotonic characteristics. ALE is normalized, at [ 0; 1] the actual amplitude is different.

Defining:

an "artificial neural network," also known as an "ANN" or simply "neural network," is based on a collection of large connected simple units called artificial neurons, which are roughly analogous to axons in the biological brain. If the combined afferent signal is strong enough, the neuron will be activated and the signal will propagate to other neurons connected to it. The activation function of such neurons is usually (although not always) expressed as a sigma function.

"deep learning" (DL) (also known as deep structured learning, hierarchical learning, or deep machine learning) is a study of artificial neural networks that include multiple hidden layers of neurons. Such neural networks are referred to as "deep neural networks". A "convolutional neural network" is a neural network in which the connectivity pattern is inspired by the animal's visual cortical tissue.

"depth feature selection" (DFS) is a machine learning method based on adding a sparse, one-to-one linear layer between the input layer and the first hidden layer of a multi-layer perceptron (MLP).

The "mean absolute error" (MAE) is a quality indicator that is equal to the mean absolute difference between the predicted and observed ages.

"exchange feature importance" (PFI) is a method of estimating feature importance in machine learning models based on measuring the reduction in predictive capability after exchanging particular features.

The 'multilayer perceptron' (MLP) is a neural network model for supervised learning, which is composed of more than 3 layers and is used for distinguishing linear indivisible data.

"whole genome sequencing" (WGS) is a comprehensive method of analyzing the entire genome, suggesting fragmentation of long DNA molecules prior to their reading. The output of the WGS method is typically a file containing 105-106 short DNA sequences.

"XGBoost" (XGB) is an open source implementation of the gradient lifting machine learning technique, relying on iterative combination with weak prediction models.

The "human intestinal microbiome" is an aggregate of all microflora that inhabit the human intestine. The microbiota is the sum of all organisms that inhabit the human gut. The term "gut microbiome sample" is generally used to refer to a "stool sample" in microbiology, and biopsy samples may also be used.

"taxonomic profiling" is a vector that includes the relative or absolute number, abundance, or percentage of all microorganisms in a community. The sum of all taxonomic profiles is 1 or 100%. I.e. the number of different microorganisms is a fraction, the sum being 1 or up to 100%.

"cumulative local effect" (ALE) is a method of visualizing the importance of features in a model. The ALE score in this patent is the predicted average change in age after replacing the abundance of the target taxon with lower and higher 5% quantiles.

A "metagenome" is the collection of all DNA information present in a population, usually represented by a text file containing short sequences (reads 100-200 nucleotides in length) determined during sequencing.

"binning" is the process of assigning each read in the sequencing output file to one or more taxa or taxa.

"phenotypic age" is a comprehensive indicator of human health, reflecting the likelihood of an individual developing age-related disease, as well as the performance of its organism, compared to healthy individuals of all chronological ages.

A "workflow" (Pipeline) is a series of sequentially executed programs that receive raw data, perform data formatting, filtering and normalization, perform calculations, and generate outputs, which may be predictive models, generative models, data statistics, or estimated hidden variables.

Chronological age (chronologic) is the actual age of a subject or organism. For animals and humans, chronological age may be based on the age calculated from conception or on the age calculated from the time and date of birth. The age of a cell, tissue or organ can be determined by the age of the subject or organism from which the cell, tissue or organ was obtained, plus the time the cell, tissue or organ was placed in culture. Alternatively, in the case of cell or tissue culture, the age may be related to the total or cumulative time of culture or the number of passages.

The phenotypic age may or may not be the chronological age of the subject or organism, but is the age at which the body or organ is biologically apparent. While a subject may actually be 40 years old, its biologically phenotypic age may be well above or below 40 years old. The phenotypic age may include a set of observable features resulting from the interaction of the individual's genotype with the environment. Phenotypic age is related to the physiological health of an individual and its biomarkers. Phenotypic age is related to how and to what extent the performance of human organs and regulatory systems maintain the overall balance of the organism's levels, as these functions generally decline with time and age.

Biological age is defined as the score of an individual that indicates the risk of a personal health problem. In the aging clock, the most typical definition is the projection of aging biomarker f (b) - > Rn on the digit line, representing chronological age. (b) constructed with a maximized probability P (CA ═ f (b) | b), where CA represents the actual chronological age of the individual; thus, the distance between the comparison f (b) and the CA conveys information about the health of the individual. This is the way in which the age of the organism is defined in the embodiment of the microbiological clock. However, biological age may also provide a risk of developing a particular health condition in the future, a remaining life expectancy, or a distance between those corresponding to age.

For the processes and methods disclosed herein, the operations performed in the processes and methods may be performed in a different order. Moreover, the outlined operations are only provided as examples, and some operations may be selective, eliminated, supplemented with further operations, or expanded into additional operations without departing from the spirit of the disclosed embodiments.

The subject may be a human, a mammal, an animal, a plant, or any multicellular organism. Examples of suitable mammals include, but are not limited to, humans, monkeys, apes, dogs, cats, cows, horses, goats, pigs, rabbits, mice, and rats.

The biological sample may be blood, lymphocytes, monocytes, neutrophils, basophils, eosinophils, myeloid cells, lymphoid cells, bone marrow, saliva, oral tissue, nasal tissue, urine, fecal material, hair, breast tissue, ovarian tissue, uterine tissue, cervical tissue, prostate tissue, testicular tissue, brain tissue, neuronal cells, astrocytes, liver tissue, kidney, thyroid tissue, stomach tissue, intestinal tissue. Pancreatic tissue, vascular tissue, skin, lung tissue, bone tissue, cartilage, ligaments, tendons, adipocytes, muscle cells, neurons, astrocytes, cultured cells of varying channel numbers, cancer or tumor cells, cancer or tumor tissue, normal cells, normal tissue, any tissue or cell containing a nucleus of genetic material, or genetic material in the form of DNA of a known or unknown subject, such as a microbial marker as described herein.

The present disclosure is not limited to the particular embodiments described in this application, which are intended to describe each aspect. Many modifications and variations are possible without departing from the spirit and scope thereof. Functionally equivalent methods and apparatuses within the scope of the present application, in addition to those enumerated herein, may be derived from the foregoing description. Such modifications and variations are intended to fall within the scope of the appended claims. The disclosure is to be limited only by the terms of the appended patent claims, along with the full scope of equivalents to which such claims are entitled. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting.

In one embodiment, the method of the present invention may comprise aspects that are executed on a computing system. Thus, a computing system may include a storage device having computer-executable instructions for performing the method. The computer-executable instructions may be part of a computer program product comprising one or more algorithms for performing any of the methods in any claim.

In an embodiment, any of the operations, processes, or methods described herein may be performed or caused to be performed in response to execution of computer readable instructions stored on a computer readable medium and executable by one or more processors. The computer-readable instructions may be executed by processors of a wide range of computing systems, from desktop computing systems, portable computing systems, tablet computing systems, handheld computing systems, and network elements and/or any other computing device. The computer readable medium is not transitory. A computer-readable medium is a physical medium having computer-readable instructions stored therein for physical reading by a computer or processor from the physical medium.

Various vehicles by which processes and/or systems and/or other technologies described herein may be effected (e.g., hardware, software, and/or firmware), and that the preferred vehicle will vary with the environment in which the processes and/or systems and/or other technologies are deployed. For example, if an implementer determines that speed and accuracy are paramount, the implementer may opt for a mainly hardware and/or firmware vehicle; if the flexibility is extremely important, an implementer can choose to mainly adopt software to realize; alternatively, the implementer may opt for some combination of hardware, software, and/or firmware.

The various operations described herein may be implemented individually and/or collectively by various hardware, software, firmware, or virtually any combination thereof. In certain embodiments, portions of the subject matter described herein may be implemented in an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA), a Digital Signal Processor (DSP), or other integrated format. However, it is possible that some aspects of the embodiments disclosed herein, in whole or in part, may be equivalently implemented in integrated circuits, as one or more computer programs running on one or more computers (e.g., as one or more programs running on one or more computer systems), as one or more programs running on one or more processors (e.g., as one or more programs running on one or more microprocessors), as firmware, or as virtually any combination thereof, and that designing the circuitry and/or writing the code for the software and/or firmware would be feasible in accordance with this disclosure. Moreover, the subject matter described herein is capable of being distributed as a program product in a variety of forms, and illustrative embodiments of the subject matter described herein apply regardless of the particular type of signal bearing media actually used. Examples of physical signal bearing media include, but are not limited to, the following: recordable type media such as floppy disks, Hard Disk Drives (HDD), Compact Disks (CD), Digital Versatile Disks (DVD), digital tape, computer memory, or any other non-transitory or transmissive physical medium. Examples of physical media having computer-readable instructions omit transitory or transmission-type media such as digital and/or analog communication media (e.g., fiber optic cables, waveguides, wired communications links, wireless communications links, etc.).

It is common practice to describe devices and/or processes in the manner described herein, and thereafter use engineering practices to integrate such described devices and/or processes into data processing systems. That is, at least a portion of the devices and/or processes described herein may be integrated into a data processing system through reasonable experimentation. A typical data processing system generally includes one or more system unit housings, a video display device, memory, such as volatile and non-volatile memory, processors, such as microprocessors and digital signal processors, computational entities, such as operating systems, drivers, graphical user interfaces and applications programs, one or more interactive devices, such as touch pads or screens, and/or a control system, including feedback loops and control motors (e.g., feedback for sensing position and/or velocity; control motors for moving and/or adjusting components and/or quantities). A typical data processing system may be implemented using any suitable commercially available components, such as those found in data computing, communication, and/or network computing or communication systems.

The subject matter described herein sometimes illustrates components contained within, or connected with, various other components. This described architecture is merely exemplary, and in fact many other architectures can be implemented which achieve the same functionality. Conceptually, any arrangement of components that achieves the same functionality is effectively "associated" such that the desired functionality is achieved. Hence, any two components herein combined to achieve a particular functionality can be seen as "associated with" each other such that the desired functionality is achieved, irrespective of architectures or intermedial components. Likewise, any two components so associated can also be viewed as being "operably connected," or "operably coupled," to each other to achieve the desired functionality, and any two components capable of being so associated can also be viewed as being "operably couplable," to each other to achieve the desired functionality. Specific examples of operative couplings include, but are not limited to: components that may physically cooperate and/or physically interact with each other, and/or components that may wirelessly interact and/or wirelessly interact with each other, and/or components that may logically interact and/or logically interact with each other.

Fig. 6 shows an exemplary computing device 600 (e.g., a computer) that may be arranged to perform the methods (or portions thereof) described herein in some embodiments. In a very basic configuration 602, computing device 600 typically includes one or more processors 604 and a system memory 606. A memory bus 608 may be used for communicating between the processor 604 and the system memory 606.

Depending on the desired configuration, the processor 604 may be of any type, including but not limited to: a microprocessor (μ P), a microcontroller (μ C), a Digital Signal Processor (DSP), or any combination thereof. The processor 604 may include one or more levels of cache, such as a level one cache 610 and a level two cache 612, a processor core 614, and registers 616. An exemplary processor core 614 may include an Arithmetic Logic Unit (ALU), a Floating Point Unit (FPU), a digital signal processing core (DSP core), or any combination thereof. An exemplary memory controller 618 may also be used with processor 604, or in some embodiments memory controller 618 may be an internal part of processor 604.

Depending on the desired configuration, system memory 606 may be of any type, including but not limited to: volatile memory (such as RAM), non-volatile memory (such as ROM, flash memory, etc.), or any combination thereof. System memory 606 may include an operating system 620, one or more application programs 622, and program data 624. Application 622 may include a determination application 626 arranged to perform operations described herein, including operations relating to methods described herein. Determination application 626 may obtain data such as pressure, flow rate, and/or temperature and determine changes to the system to change the pressure, flow rate, and/or temperature.

Computing device 600 may have additional features or functionality, and additional interfaces to facilitate communications between basic configuration 602 and any required devices and interfaces. For example, a bus or interface controller 630 may be used to facilitate communication between basic configuration 602 and one or more data storage devices 632 via a storage interface bus 634. The data storage device 632 may be a removable storage device 636, a non-removable storage device 638, or a combination thereof. Removable storage and non-removable storage devices include, to name a few: magnetic disk devices, such as flexible disk drives and Hard Disk Drives (HDDs); optical disk drives, such as Compact Disk (CD) drives or Digital Versatile Disk (DVD) drives, Solid State Drives (SSDs), and tape drives. Example computer storage media may include: volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data.

The system memory 606, removable storage 636 and non-removable storage 638 are examples of computer storage media. Computer storage media include, but are not limited to: computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, Digital Versatile Disks (DVD) or other optical storage, magnetic tape, magneto-optical disk storage or other magnetic storage devices, or any other medium which can be used to store the desired information and which can accessed by computing device 600. Any such computer storage media may be part of computing device 600.

Computing device 600 may also include an interface bus 640 for facilitating communication from various interface devices (e.g., output devices 642, peripheral interfaces 644, and communication devices 646) to the basic configuration 602 via the bus or interface controller 630. Exemplary output devices 642 include a graphics processing unit 648 and an audio processing unit 650, which may be configured to communicate with various external devices such as a display or speakers via one or more A/V ports 652. Exemplary peripheral interfaces 644 include a serial interface controller 654 or a parallel interface controller 656, which may be configured to communicate with external devices such as input devices (e.g., keyboard, mouse, pen, voice input device, touch input device, etc.) or other peripheral devices (e.g., printer, scanner, etc.) via one or more I/O ports 658. The exemplary communication device 646 includes a network controller 660, which may be arranged to facilitate communications with one or more other computing devices 662 over a network communication link through one or more communication ports 664.

A network communication link may be one example of a communication medium. Communication media may typically be embodied by computer readable instructions, data structures, program modules, or other data in a modulated data signal, such as a carrier wave or other transport mechanism, and may include any information delivery media. A "modulated data signal" may be a signal that has one or more of its characteristics set or changed in such a manner as to encode information in the information. By way of example, and not limitation, communication media may include wired media such as a wired network or direct-wired connection, and wireless media such as acoustic, Radio Frequency (RF), microwave, Infrared (IR), and other wireless media. The term computer readable media as used herein may include both storage media and communication media.

Computing device 600 may be implemented as a portion of a small-form factor portable (or mobile) electronic device such as a cell phone, a Personal Data Assistant (PDA), a personal media player device, a wireless network watch device, a personal headset device, an application specific device, or a hybrid device that include any of the above functions. Computing device 600 may also be implemented as a personal computer, including both notebook and non-notebook configurations. Computing device 600 may also be any type of network computing device. Computing device 600 may also be an automated system as described herein.

The embodiments described herein may comprise a special purpose or general-purpose computer including various computer hardware or software modules.

Embodiments within the scope of the present invention also include computer-readable media for carrying or having computer-executable instructions or stored data structures. Such computer-readable media can be any available media that can be accessed by a general purpose or special purpose computer. By way of example, and not limitation, such computer-readable media can comprise RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to carry or store desired program code means in the form of computer-executable instructions or data structures and which can be accessed by a general purpose or special purpose computer. When information is transferred or provided over a network or another communications connection (either hardwired, wireless, or a combination of hardwired or wireless) to a computer, the computer properly views the connection as a computer-readable medium. Thus, any such connection is properly termed a computer-readable medium. Combinations of the above should also be included within the scope of computer-readable media.

Computer-executable instructions comprise, for example, instructions and data which cause a general purpose computer, special purpose computer, or special purpose processing device to perform a certain function or group of functions. Although the subject matter has been described in language specific to structural features and/or methodological acts, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are disclosed as example forms of implementing the claims.

With respect to the use of substantially any plural and/or singular terms herein, those having skill in the art can translate from the plural to the singular and/or from the singular to the plural as is appropriate to the context and/or application. Various combinations of singular (plural) numbers may be explicitly set forth herein for clarity.

It will be understood by those within the art that, in general, terms used herein in reference to appended claims (e.g., such as the subject matter of the appended claims) are generally intended as "open" terms (e.g., the term "including" should be interpreted as "including but not limited to," the term "having" should be interpreted as "having at least," the term "includes" should be interpreted as "includes but is not limited to," and the like). It will be further understood by those within the art that if a specific number of an introduced claim recitation is intended, such an intent will be explicitly recited in the claim; in the absence of such a statement, there is no such intent. For example, as an aid to understanding, the following appended claims may contain usage of introductory phrases such as "at least one" and "one or more" to introduce claim recitations. However, the use of such phrases should not be construed to imply that the introduction of a claim recitation by the indefinite articles "a" or "an" limits any particular claim containing such introduced claim recitation to embodiments containing only one such recitation, even when the same claim includes the introductory phrases "one or more" or "at least one" and indefinite articles such as "a" or "an" (e.g., "a" and "an" should be interpreted to mean "at least" one "or" one or more "); the same holds true for the use of definite articles used to introduce claim recitations. Furthermore, even if a specific number of an introduced claim recitation is explicitly recited, those skilled in the art will recognize that such recitation should be interpreted to mean at least the recited number (e.g., the bare recitation of "two recitations," without other modifiers, means at least two recitations, or two or more recitations). Further, where the use of "at least one of A, B and C, etc." is used in general, it is intended that the word "structure" be used in the sense one of ordinary skill in the art would understand (e.g., "having at least one of A, B and C" would include but not be limited to systems having A alone, B alone, C, A and B together alone, A and C together, B and C together, and/or A, B and C together, etc.). It will be further understood by those within the art that virtually any disjunctive word and/or phrase presenting two or more alternative terms, whether in the description, claims, or drawings, should be understood to contemplate the possibilities of including one, the other, or both. For example, the phrase "a or B" will be understood to encompass the possibility of "a" or "B" or "a and B".

Furthermore, where features or aspects of the disclosure are described in terms of Markush groups, those skilled in the art will recognize that the disclosure is also thereby described in terms of any individual member or subgroup of members of the Markush group.

As will be understood by those skilled in the art, for any and all purposes, such as in providing a written description, all ranges disclosed herein can also include any and all possible subranges and combinations of subranges thereof. Any listed range can be easily identified as being fully described, and such that the same range is broken down into at least equal halves, thirds, quarters, fifths, tenths, etc. As non-limiting examples, each of the ranges discussed herein can be readily subdivided into a lower third, a middle third, and an upper third, to name a few. It will also be understood by those within the art that all terms, such as "at most," "at least," and the like, are to be interpreted as encompassing the recited number, and are to be interpreted as being subsequently subdivided into the aforementioned subranges. Finally, as understood by those of skill in the art, a range encompasses each individual member. Thus, for example, a population of 1 to 3 cells refers to a population of 1, 2 or 3 cells. Similarly, a group of 1 to 5 cells refers to a group of 1, 2, 3, 4, or 5 cells, and so forth.

From the foregoing, it will be appreciated that various embodiments of the disclosure have been described herein for purposes of illustration, and that various modifications may be made without deviating from the scope and spirit of the disclosure. Accordingly, the various embodiments disclosed herein are not intended to be limiting, with the true scope and spirit being indicated by the following claims.

Cross-reference to the present patent application: us 16/415,855 filed on 17.5.2019; us 16/104,391 filed on 8/17/2018; us 16/044,784 filed on 25.7.2018; us 62/536,658 filed on 25.7.7.2017; and us 62/547,061 filed on 2017, 8, 17, each of which is incorporated herein by specific reference.

All references mentioned herein are incorporated herein by specific reference.

Reference to the literature

Blalcock, Eric m., Kuey-Chu Chen, Keith Sharrow, James p.herman, Nada m.porter, Thomas c.foster, and Philip w.landfield. In 2003. Hippocampus area aged gene microarray: statistical analysis identified a new process related to cognitive impairment (journal of neuroscience: official journal of neuroscience institute 23(9): 3807-19.

Chovers, Itay, Dongmei Liu, Ronald H.Farkas, Tushara L.Gunatilaka, Abigail S.Hackam, Steven L.Bernstein, Peter A.Campachhiaro, Giovanni Parmigiani, and Donald J.Zack. In 2003. 2881-93 in human molecular genetics 12(22) of changes in gene expression in adult human retina.

Hong, mu-Gwan, Amanda j.myers, patrick k.e.magnusson, and Jonathan a.prince. In 2008. Transcriptome assessment of human brain and lymphocyte senescence. PloS One 3(8) e 3024.

Horvath, Steve. 2013. Age of DNA methylation of human tissues and cell types genome biology 14(10) R115.

Horvath, Steve, yang Zhang, Peter Langfelder, Ren s.kahn, Marco p.m.boks, kritel van Eijk, Leonard h.van den Berg, and Roel a.operoff. 2012. Influence of senescence on DNA methylation modules in human brain and blood tissues genome biology 13(10) R97.

Pedro de、Curdo and George m.church. In 2009. Meta-analysis of age-related gene expression profiling identified common features of senescence, bioinformatics 25(7), 875-81.

Mendelsohn, Andrew r. and James w.larrick. 2013. DNA methylsets as biomarkers of epigenetic instability and human aging-Return to the youth research 16(1) 74-77.

Park, Sang-Kyu, Kyoungmi Kim, Grier P.Page, David B.Allison, Richard Weinruch and Tomas A.Prolla. In 2009. Gene expression profiling of Multi-lineage mice aging: the influence of aging biomarkers and dietary antioxidants is recognized in aging cells 8(4) 484-95.

Park, Sang-Kyu and Tomas A.Prolla. In 2005. Study on Gene expression profiling of myocardial and skeletal muscle aging in cardiovascular studies 66(2) and 205-12.

Weinruch, Richard, Tsuyoshi Kayo, Cheol-Koo Lee, and Tomas A.Prolla. In 2002. Gene expression profiling for senescence Using DNA microarrays (mechanism of senescence and progression) 123(2-3) 177-93.

Welle, Stephen, Andrew i.brooks, Joseph m.delehanty, Nancy needle and Charles a.thornton. In 2003. 149-59 in physiological genomics 14(2), analysis of human muscle senescence genes.

Wesolowska-Andersen, Agata, Martin Iain Bahl, Vera Carvalho, Karsten Kristiansen, Thomas Sicheritz-Ponten, Ramnek Gupta, and Tine Rask Licht. 2014. The choice of methods for extracting bacterial DNA from feces has influenced the community structure evaluated by metagenomic analysis, microbiology 2 (June): 19.

wolters, Stefanie andschumacher. 2013. Genome maintenance and transcriptional integrity in aging and disease 4 frontier in genetics. https:// doi.org/10.3389/fgene.2013.00019.

Zahn, Jacob M., Suresh Poosala, Art B.Owen, Donald K.Ingram, Ana Lustig, Arnell Carter, Ashani T.Weerarnana, and the like. In 2007. Agamp: gene expression database of mouse senescence. (e 201) in "PLoS genetics" 3 (11).