阅读说明：本技术 一种提高宏基因组纳米孔测序数据菌种组装效率的方法 (Method for improving strain assembly efficiency of metagenome nanopore sequencing data ) 是由 李振中 陈莉 李珊 戴岩 李诗濛 任用 于 2021-08-13 设计创作，主要内容包括：本发明提供一种通过降维聚类分群提高宏基因组测序数据菌种组装效率的方法,所述方法通过k-mer频率或频数统计,在组装前进行降维预分群,能够显著提高宏基因组组装效率,组装时间至少减少一半以上,同时保证生信鉴定的有效性和准确性。(The invention provides a method for improving the assembly efficiency of a metagenome sequencing data strain through dimension reduction clustering, which is characterized in that dimension reduction pre-clustering is carried out before assembly through k-mer frequency or frequency statistics, the assembly efficiency of the metagenome can be obviously improved, the assembly time is reduced by more than half at least, and the validity and the accuracy of letter identification are ensured at the same time.)

1. A method for improving the assembly efficiency of metagenome sequencing data strains by dimensionality reduction and grouping is characterized by comprising the following steps:

step 1) sequence generation: generating fastq format sequence information by using the metagenome sequencing off-line data;

step 2) sample splitting: carrying out sample splitting according to the library tag sequence;

step 3) sequence quality control: including but not limited to quality control of sequence length and/or quality;

step 4), calculating a k-mer frequency or frequency matrix: performing k-mer frequency or frequency matrix calculation based on the sequence;

step 5), dimension reduction clustering grouping processing: performing dimensionality reduction clustering grouping processing on all the sequences based on the frequency matrix;

step 6) sequence assembly: and respectively assembling the sequences of each cluster after dimension reduction clustering.

2. The method for improving the assembling efficiency of the metagenome sequencing data strain by dimension reduction and grouping according to claim 1, wherein in the step 4), k is 2-20000; preferably, k is 5 to 75.

3. The method for improving the assembling efficiency of metagenomic sequencing data strains through dimension reduction clustering as claimed in any one of claims 1-2, wherein the dimension reduction clustering of step 5) is performed by using methods including but not limited to: carrying out sequence dimension reduction clustering grouping on the Umap, the t-SNE and the KNN;

preferably, the sequences after dimension reduction and clustering in step 5) are further processed by polish respectively.

4. The method for improving the assembling efficiency of the metagenome sequencing data strains through dimension reduction clustering as claimed in any one of claims 1 to 3, wherein the assembling in the step 6) is that the sequences of each cluster after each clustering of poll are assembled respectively;

preferably, the assembly uses include, but are not limited to: Canu/meta-Flye, wtbg 2, NECAT software.

5. The method for improving the assembling efficiency of the metagenome sequencing data strains through dimensionality reduction, clustering and grouping according to any one of claims 1 to 4, wherein the samples in the step 2) are split into: the sequence is split into sequence sets belonging to different samples according to the tag sequences of the library, and the linker sequences are removed at the same time.

6. The method for improving the assembly efficiency of the metagenomic nanopore sequencing data strain by dimension reduction, clustering and grouping according to any one of claims 1 to 5, wherein the step 3) further comprises a sequence correction step after the sequence quality control: and (4) self-correcting the filtered sequence to correct the base with sequencing error.

7. The method for improving the strain assembly efficiency of metagenomic sequencing data by dimensionality reduction and clustering according to any one of claims 1 to 7, wherein the sequencing data is first generation, second generation, third generation or fourth generation sequencing data;

preferably, the sequencing data is four generation nanopore sequencing data.

8. A method of generating a credit for identification of a species, the method comprising the method of any one of claims 1 to 7 and further comprising:

step 7) species identification: species identification was performed based on the assembled sequences.

9. A species identification device, comprising: at least one memory for storing a program; at least one processor configured to load the program to perform the method of any of claims 1-8.

10. A storage medium having stored therein processor-executable instructions, which when executed by a processor, are configured to implement the method of any one of claims 1-8.

Technical Field

The invention relates to the field of biogenic analysis, in particular to a method for improving the strain assembly efficiency of metagenome nanopore sequencing data through dimensionality reduction.

Background

Metagenomics (also known as Metagenomics) is the study of genomics of microorganisms in their original place of life. The metagenome directly extracts DNA or RNA of all microorganisms from an environment sample, constructs a metagenome library and sequences, and systematically analyzes genetic diversity and functional diversity of the microorganisms in the environment to explore the fields of taxonomy, function, evolution and the like. Metagenomics allows us to directly investigate the genetic composition of microbial communities such as bacteria, viruses and fungi, beyond the limitations of culturability and taxonomic properties. The analysis content of the metagenomics mainly comprises species composition and difference analysis, functional composition and difference analysis of microbial communities, relationship between environmental factors and microbial communities and the like.

Nanopore sequencing technology (also known as fourth generation sequencing technology) is a new generation of sequencing technology that has emerged in recent years. Currently the sequencing length can reach 150 kb. This technology began in the 90 s and underwent three major technological innovations: firstly, single-molecule DNA passes through the nanopore; secondly, controlling the accuracy of sequencing molecules on single nucleotide by enzyme on the nano-pores; thirdly, controlling the sequencing precision of the mononucleotide. The widely accepted Nanopore sequencing platform on the market today is the MinION and GridION Nanopore sequencer from Oxford Nanopore Technologies (ONT for short). The method is characterized by single-molecule sequencing, and has the characteristics of long sequencing read length, convenience in library preparation, high sequencing speed, real-time acquisition of sequencing data and the like.

It is clearly the ideal case that the nanopore sequencing based metagenomics study target is total DNA in the entire habitat and in order to obtain information on the complete genome in environmental samples, the full-length genomic sequence of each microorganism needs to be restored. But utilizes the technology of the de novo assembly of metagenome, namely metagenome reads are firstly assembled into contigs, classification or phylogenetic information is assigned to each contig through sequence comparison with a reference genome to obtain species components of a microbial community, and further differential analysis, functional analysis and the like of the community are carried out.

The current assembly analysis process of nanopore sequencing data is as follows:

1) during the sequencing run, using ONT MinKNOW software to collect raw sequencing data;

2) generating a base sequence of the original data by using ONT Albacore or ONT Guppy software;

3) filtering out sequences with the length of less than 500bp and the average sequencing quality value of less than 8 by using a self-editing python script;

4) using consensus software to perform sequence self-correction;

5) sequence polish was performed using medaka software;

6) strain assembly was performed using Canu/meta-Flye software.

However, in practice, the sequencing reads have large data volume, long assembly running time and low reads utilization rate. In particular, since metagenomic sequencing is directed to all microbial sequences in a complex environment, difficulties are increased in assembly due to species diversity and high sequence similarity of closely related species, thereby increasing assembly running time.

In view of the above, the present invention is particularly proposed.

Disclosure of Invention

The invention aims to seek to improve the assembly efficiency of the metagenome nanopore sequencing data strain. In order to achieve the purpose, the invention provides a brand new idea that sequencing data is identified by a dimensionality reduction clustering pre-clustering mode before sequence assembly.

The specific technical scheme is as follows:

the invention firstly provides a method for improving the assembly efficiency of metagenome sequencing data strains by dimension reduction, clustering and grouping, which is characterized by comprising the following steps:

step 1) sequence generation: generating fastq format sequence information by using the metagenome sequencing off-line data;

step 2) sample splitting: carrying out sample splitting according to the library tag sequence;

step 3) sequence quality control: including but not limited to quality control of sequence length and/or quality;

step 4), calculating a k-mer frequency or frequency matrix: performing k-mer frequency or frequency matrix calculation based on the sequence;

step 5), dimension reduction clustering grouping processing: performing dimensionality reduction clustering grouping processing on all sequencing sequences based on the frequency matrix;

step 6) sequence assembly: and respectively assembling the sequences of each cluster after dimensionality reduction and clustering.

Further, the step 2) of splitting the sample into: the sequence is split into sets of sequences belonging to different samples according to the tag sequence of the library (e.g., barcode), while removing the linker sequence.

Further, the sequence quality control of the step 3) is as follows: counting the length and quality values of the sequence;

in some embodiments, such as for nanopore length read data, filter out length less than 500bp and average sequencing quality value less than 8 sequence.

Further, in the step 4), k is 2 to 20000, preferably, k is 5 to 75; more preferably, k is 5, specifically: the number of sequence types of the 5-mers is 4 × 4/2 × 512, and the frequency or frequency count of 512 mers in each reads is calculated to obtain a 5-mer frequency or frequency matrix.

Further, the dimensionality reduction clustering of the step 5) includes but is not limited to: carrying out sequence dimensionality reduction clustering on the Umap, the t-SNE and the KNN;

in some embodiments, sequence dimension reduction clustering is performed using a Umap package; the parameters of the dimensionality reduction clustering are set as follows: random _ state 42, n _ neighbors 30, min _ dist 0.0, n _ components 2; then clustering is carried out by using a hdbscan packet of python according to the result of Umap dimension reduction clustering, and each read is determined to belong to a cluster.

Further, the reads sequence after the dimensionality reduction, clustering and clustering in the step 5) can further comprise respective polish processing; preferably, for each cluster's reads, the polish process is performed separately using medaka software.

Further, the assembling in the step 6) is to assemble the reads of each cluster which is subjected to poll respectively;

in some embodiments, the assembly uses include, but are not limited to: Canu/meta-Flye, wtbg 2, NECAT software.

The invention also provides a method for generating a credit for identifying species, which is characterized by comprising the method and further comprising the following steps: step 9) species identification: species identification was performed based on the assembled sequences.

The present invention also provides a biological information analysis device for identifying a species, comprising: at least one memory for storing a program; at least one processor configured to load the program to perform the method as described above.

The present invention also provides a storage medium having stored therein processor-executable instructions for implementing the method as described above when executed by a processor.

Further, the sequencing data is first generation, second generation, third generation or fourth generation sequencing data; preferably, four generation nanopore sequencing data.

The invention has the beneficial technical effects that:

according to the invention, the metagenome data is classified into different cluster according to strains through dimensionality reduction, clustering and grouping, and then each cluster is assembled respectively, so that the metagenome assembly efficiency can be remarkably improved, the assembly time is reduced by at least over half, and the metagenome data is consistent with species identification results of non-clustered assembly.

The invention effectively improves the identification efficiency of the metagenome and simultaneously ensures the effectiveness and the accuracy of strain identification.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.

FIG. 1 is a flow chart of the method of the present invention;

FIG. 2 is a graph of Umap dimension reduction clustering results of 1h in example 2; measuring 1h output data on a four-generation nanopore sequencing platform, and obtaining a clustering result according to Umap dimension reduction clustering, wherein each point clustered together is the same clustering;

FIG. 3 is a graph of the 2h Umap dimension reduction clustering result in example 2; 2h output data measured on a four-generation nanopore sequencing platform are clustered according to a clustering result obtained by Umap dimension reduction in the invention, wherein each point gathered together is the same cluster;

FIG. 4 is a graph of the 3h frequency Umap dimension reduction clustering result in embodiment 2; 3h output data measured by a four-generation nanopore sequencing platform are clustered according to the clustering result obtained by Umap dimension reduction clustering, wherein each point clustered together is the same cluster;

FIG. 5 is a graph of the Umap dimension-reduction clustering result of 4h in embodiment 2; 4h output data measured on a four-generation nanopore sequencing platform are clustered according to the clustering result obtained by Umap dimension reduction clustering, wherein each point clustered together is the same cluster;

FIG. 6 is a diagram showing the result of Umap dimension-reducing clustering in 5h in example 2; 5h output data measured on a four-generation nanopore sequencing platform are clustered according to the clustering result obtained by Umap dimension reduction clustering in the invention, wherein each point clustered together is the same cluster.

Detailed Description

Embodiments of the present invention will be described in detail below with reference to examples, but those skilled in the art will appreciate that the following examples are only illustrative of the present invention and should not be construed as limiting the scope of the present invention, and that the examples are a part of, but not all of the examples of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

Definition of partial terms

Unless defined otherwise below, all technical and scientific terms used in the detailed description of the present invention are intended to have the same meaning as commonly understood by one of ordinary skill in the art. While the following terms are believed to be well understood by those skilled in the art, the following definitions are set forth to better explain the present invention.

As used herein, the terms "comprising," "including," "having," "containing," or "involving" are inclusive or open-ended and do not exclude additional unrecited elements or method steps. The term "consisting of …" is considered to be a preferred embodiment of the term "comprising". If in the following a certain group is defined to comprise at least a certain number of embodiments, this should also be understood as disclosing a group which preferably only consists of these embodiments.

The terms "about" and "substantially" in the present invention denote an interval of accuracy that can be understood by a person skilled in the art, which still guarantees the technical effect of the feature in question. The term generally denotes a deviation of ± 10%, preferably ± 5%, from the indicated value.

Where an indefinite or definite article is used when referring to a singular noun e.g. "a" or "an", "the", this includes a plural of that noun.

Furthermore, the terms first, second, third, (a), (b), (c), and the like in the description and in the claims, are used for distinguishing between similar elements and not necessarily for describing a sequential or chronological order. It is to be understood that the terms so used are interchangeable under appropriate circumstances and that the embodiments of the invention described herein are capable of operation in other sequences than described or illustrated herein.

The data dimension reduction clustering of the invention is that a series of related high-dimensional variables are reduced into a series of low-dimensional variables, the low-dimensional variables reflect the characteristics of original data as much as possible, and the data with similar characteristics are clustered into one class. The invention preferably adopts a UMAP dimension reduction clustering algorithm and dimension reduction clustering of a 5-mer frequency matrix based on a sequencing sequence. The UMAP (Uniform Manifold Approximation and Projection for Dimension Reduction) is a Dimension Reduction technique, similar to t-SNE, and can be used for visualization, but also for general nonlinear Dimension Reduction clustering.

Metagenomics (Metagenomics) described in this invention is a genomics study of microorganisms in their original place of life. The metagenome directly extracts DNA or RNA of all microorganisms from an environment sample, constructs a metagenome library and sequences, and systematically analyzes genetic diversity and functional diversity of the microorganisms in the environment to explore the fields of taxonomy, function, evolution and the like. The analysis content of the metagenomics mainly comprises species composition and difference analysis, functional composition and difference analysis of microbial communities, relationship between environmental factors and microbial communities and the like.

The nanopore sequencing technology (also called fourth generation sequencing technology) provided by the invention is a new generation sequencing technology which has emerged in recent years. Currently the sequencing length can reach 150 kb. The method is characterized by single-molecule sequencing, and has the characteristics of long sequencing read length, convenience in library preparation, high sequencing speed, real-time acquisition of sequencing data and the like.

The method for improving the assembling efficiency of the metagenome sequencing data strains by dimensionality reduction and grouping mainly comprises the following steps based on a pre-grouping mode: step 1) sequence generation: generating fastq format sequence information by using the metagenome sequencing off-line data; step 2) sample splitting: splitting the sequence into sequence sets belonging to different samples according to the library label sequence; step 3) sequence quality control: such as quality control of sequence length and/or quality, etc.; step 4), calculating a k-mer frequency or frequency matrix, wherein k is 2-20000; step 5), dimension reduction clustering (cluster) processing: performing dimensionality reduction grouping processing on all sequencing sequences based on the frequency matrix; step 6) sequence assembly: and respectively assembling the sequences of the clusters after dimensionality reduction.

In some aspects, said step 3) comprises, for example, the length and quality values of the statistical sequence. Illustratively, for nanopore sequencing data, sequences with a length of less than 500bp and an average sequencing quality value of less than 8 were filtered out. The quality control standard can be appropriately selected in the art according to the actual sequencing data.

In some aspects, the step 3) may further include, after the sequence quality control, a sequence correction step: and (4) self-correcting the filtered sequence to correct the base with sequencing error.

In some aspects, k in the step 4) is 2 to 20000, preferably k is 5 to 75.

It can be understood that the k-mer in the invention is a subsequence with a length of k in a biological sequence, and the value of k in the method of the invention can be any positive integer, which is allowed only when the k-mer frequency can be calculated, so the value of k can be a positive integer greater than 2 in principle, and when the limitation of the actual sequence length is considered, the preferable value of k is 2-20000; more preferably 5 to 75.

In some specific examples, taking k as 5 as an example, the sequence number of 5-mers is 4 × 4/2 × 512, and the frequency or frequency count of 512 mers in each reads is calculated to obtain a 5-mer frequency or frequency matrix.

In some aspects, the reads after the dimensionality reduction clustering of step 5) are further subjected to a polich process respectively, for example, the polich process is performed respectively by using medaka software.

In some specific examples, the dimension reduction clustering of step 5) includes, but is not limited to: carrying out sequence dimensionality reduction clustering on the Umap, the t-SNE and the KNN; the different dimension reduction algorithms can perform clustering operation and do not influence the core of the invention.

Taking Umap as an example, the parameters for dimension reduction are set as follows: random _ state 42, n _ neighbors 30, min _ dist 0.0, n _ components 2; and then clustering according to the Umap dimension reduction result and determining that each read belongs to a cluster.

In some aspects, the assembling in step 6) is to assemble the reads of each poll cluster which has done poll separately;

in some specific examples, the assembly uses include, but are not limited to: Canu/meta-Flye, wtbg 2, NECAT software; the different dimension reduction clustering algorithms can carry out clustering and do not influence the core of the invention.

The sequence is split into sequence sets belonging to different samples according to the tag sequences of the library, and the linker sequences are removed at the same time.

It can be understood that the core idea of the present invention is not limited to the sequencing platform, because the calculation of k-mer frequency or frequency number for the sequence is not limited by the sequencing platform, and therefore, the sequencing data applicable to the dimension-reduced clustering assembly method of the present invention includes first generation, second generation, third generation or fourth generation sequencing data; preferably, the sequencing data is four generation nanopore sequencing data.

EXAMPLE 1 construction of the patented Process

The focus of this patent lies in, behind the metagenome data pre-clustering, assembles the promotion equipment efficiency based on the reads after the clustering.

Method optimization process

Two aspects need to be explained first: from the reads sequence to the 5-mer frequency matrix, and the cluster label obtained for each read.

In the specific calculation, the calculation is carried out,

1. first, a 5-mer frequency matrix is calculated based on the reads sequence:

-5-mers with sequence number 4 × 4/2 ═ 512;

-calculating the frequencies of the 512 5-mers in each reads;

-obtaining a 5-mer frequency matrix;

2. then, dimension reduction is carried out by using Umap based on the frequency matrix, and each reads is allocated with a cluster label by using hdbscan.

3. Then assembled for each cluster using Canu/meta-Flye software.

4. And finally, comparing the assembly result with a blast and nt database to identify species.

The invention selects ZymoBIOMICS^TMSequencing data of official ONTs of the Microbial Community DNA Standard (known in species, 8 bacteria and 2 fungi) are obtained, the sequencing data of the first 5 hours are selected according to the sequencing time and are respectively the off-machine data of sequencing 1h, 2h, 3h, 4h and 5h, and the base data amount is 458M, 919M, 1.3G, 1.7G and 2.2G. The influence of the dimensionality reduction grouping on the assembly efficiency and the accuracy of strain identification under the conditions of different time points and different data volumes is verified aiming at the sequences of the 5 time points.

And testing Canu software to directly assemble the time and strain identification results of all reads, and comparing the time and strain identification results with the assembly time and strain identification results respectively assembled by Canu after dimension reduction clustering.

Secondly, the analysis and identification process of the invention is established as follows:

1. and (3) sequence generation: and converting the electric signal into a base signal through ONT gummy software to obtain the sequence information in the fastq format from the data generated by the ONT Gridios sequencing platform.

2. Splitting a sample: using ONT Guppy software, the sequences were split into sets of sequences belonging to different samples according to the barcode sequences of the library, while removing the linker sequences.

3. And (3) sequence quality control: the length and quality values (quality score) of the sequences were counted, and for the Nanopore long read length data, sequences with a length of less than 500bp or an average sequencing quality value of less than 8 were filtered out.

4. And (3) sequence correction: the filtered sequence was self-corrected using consensus software to correct the base with sequencing error.

5. Frequency matrix: 512 5-mer frequency matrices were calculated using python script.

Umap dimensionality reduction: sequence dimension reduction clustering was performed using the Umap package of python. The parameters are set as follows: random _ state 42, n _ neighbors 30, min _ dist 0.0, n _ components 2.

Hdbscan determine cluster: and determining that each read belongs to a cluster according to the Umap dimension reduction result by using the hdbscan packet of python.

8. Assembling: for each cluster's reads, assemble separately using Canu/meta-Flye software.

9. Species identification: and comparing the assembled contig sequence with the nt library to obtain a species identification result.

Example 2 Umap clustering effect of the patented method

According to the invention, sequencing data of zymo official ONT are clustered under different time/data volume gradients in a pre-clustering-based mode, reads from the same species tend to be classified into the same cluster, and the specific implementation mode is carried out based on the flow of example 1.

The dimension reduction clustering results after the Umap clustering are shown in fig. 2-6, fig. 2 is a dimension reduction clustering result graph of 1h, fig. 3 is a dimension reduction clustering result graph of 2h, fig. 4 is a dimension reduction clustering result graph of 3h, fig. 5 is a dimension reduction clustering result graph of 4h, and fig. 6 is a dimension reduction clustering result graph of 5 h. It can be seen that all reads are sorted into different cluster by pre-clustering.

Example 3 this patented method Assembly efficiency assessment

According to the invention, the sequencing data of the zymo official ONT is obviously improved under different time/data volume gradients, such as the base data volume of 1 h-5 h, by means of pre-clustering. The detailed description is based on the procedure of example 1.

Assembly time results table 1, it can be seen that the assembly time is nearly halved using the Umap pre-divide group.

TABLE 1

Time	base(bp)	Assembly (no _ Umap)	Assembly time (Umap)
				1h	458,473,600	45m47.655s	14m36.602s
2h	919,961,649	503m13.250s	36m54.974s
				3h	1,375,306,551	749m23.833s	65m43.655s
4h	1,796,485,159	1126m10.946s	154m36.229s
				5h	2,205,881,698	1359m9.468s	179m0.873s

EXAMPLE 4 effectiveness and accuracy of the patented method

The invention uses official data of the zymo, and the species contained in the data is known, so the accuracy of assembly and species identification can be verified by comparing the species identification result with the zymo species. To verify the accuracy of the species identification of the Umap pre-clustering, we compared the clustered assembled sequences with the nt library and the species identification results of the directly assembled sequences.

The results of the identification of the strains are shown in Table 2 (taking the results of the 1h off-line data as an example), it can be seen that the identification of the strains is basically consistent, and the identified species is completely consistent with that of zymo, which fully proves the effectiveness and accuracy of the method of the present invention.

Table 21 h offline data species identification results

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present invention.