Method for improving LC-MS data signal-to-noise ratio
阅读说明：本技术 提高lc-ms数据信噪比的方法 (Method for improving LC-MS data signal-to-noise ratio ) 是由 李文杰 俞晓峰 李锐 韩双来 杨继伟 郭子杨 于 2020-12-31 设计创作，主要内容包括：本发明提供了提高LC-MS数据信噪比的方法,所述提高LC-MS数据信噪比的方法包括步骤：(A1)获得色谱峰s(t)的开始位置w-1和结束位置w-2；(A2)将色谱峰s(t)划分为峰部分s-1(t)和非峰部分s-2(t),分别进入步骤(B1)和步骤(C1)；(B1)根据所述色谱峰s(t)设计基函数；(B2)得到峰部分s-1(t)的第一层近似系数a-1和第一层细节系数d-1；(B3)得到峰部分s-1(t)的多层近似系数a-n和各层细节系数d-1,d-2···d-n；(B4)得到峰部分s-1(t)的近似部分s-(1a)(t)和细节部分s-(1b)(t)；细节部分s-(1b)(t)进行EMD经验模态分解；(B5)计算每个模态分量的能量值(B6)得到去噪后的峰部分信息s′-1(t)＝s-(1a)(t)+s′-(1b)(t)；(C1)对于非峰部分s-2(t),按照步骤(B1)-(B4)的方式获得非峰部分s-2(t)的近似部分s-(2a)(t)；(C2)整合峰部分信息s′-1(t)和近似部分s-(2a)(t),得到去噪的色谱峰s′(t)。本发明具有去噪效果好等优点。(The invention provides a method for improving LC-MS data signal-to-noise ratio, which comprises the following steps: (A1) obtaining the start position w of the chromatographic peak s (t) 1 And an end position w 2 (ii) a (A2) Dividing chromatographic peaks s (t) into peak portions s 1 (t) and non-peak portions s 2 (t) proceeding to step (B1) and step (C1), respectively; (B1) designing a basis function according to the chromatographic peak s (t); (B2) obtaining a peak portion s 1 First layer approximation coefficient a of (t) 1 And first layer detail coefficient d 1 (ii) a (B3) Obtaining a peak portion s 1 (t) multilayer approximation coefficient a n And each layer detail coefficient d 1 ，d 2 ···d n (ii) a (B4) Obtaining a peak portion s 1 (t) approximate fraction s 1a (t) and details s 1b (t); detail section s 1b (t) performing EMD empirical mode decomposition; (B5) calculating an energy value for each modal component (B6) Obtaining denoised peak part information s' 1 (t)＝s 1a (t)+s′ 1b (t); (C1) for non-peak portions s 2 (t) obtaining non-peak portions s in the manner of steps (B1) - (B4) 2 (t) approximate fraction s 2a (t); (C2) integration peak partial information s' 1 (t) and an approximation part s 2a (t), a denoised chromatographic peak s' (t) is obtained. The invention has the advantages of good denoising effect and the like.)
1. The method for improving the signal-to-noise ratio of LC-MS data comprises the following steps:
(A1) detecting the peak position p and the width w of the chromatographic peak s (t) to obtain the starting position w of the chromatographic peak s (t)1And an end position w2；
(A2) Dividing the chromatographic peak s (t) into peak portions s1(t) and non-peak portions s2(t) proceeding to step (B1) and step (C1), respectively;
(B1) designing a basis function according to the chromatographic peak s (t), wherein the discretization of the basis function is represented as F, and four orthogonal filters with the length of L are formed according to F; L-D, H-D and L-R, H-R, where L-R is a normalized representation of F, H-R is an orthogonal inverse filter of L-R, L-D is the inverse of L-R, and H-D is the inverse of H-R;
(B2) obtaining a peak portion s1First layer approximation coefficient a of (t)1And first layer detail coefficient d1；
L is the filter length;
(B3) obtaining the peak portion s by means of the step (B2)1(t) multilayer approximation coefficient anAnd each layer detail coefficient d1，d2···dnN is an integer greater than 2;
(B4) will approximate the coefficient anAnd each layer detail coefficient d1，d2···dnRespectively reconstructing to obtain peak portions s1(t) approximate fraction s1a(t) and details s1b(t)；
Detail section s1b(t) performing EMD empirical mode decomposition,imf is the modal component at different scales, rn(t) is the residual component;
(B5) calculating an energy value for each modal componentObtaining the energy mutation occurring at the k imf component;
(B6) reconstructing low-frequency imf components behind the kth imf component and residual error items thereof to obtain detail part information after effective information is extractedThereby obtaining the denoised peak partial information s'1(t)＝s1a(t)+s′1b(t)；
(C1) For non-peak portions s2(t) obtaining non-peak portions s in the manner of steps (B1) - (B4)2(t) approximate fraction s2a(t) and the details;
(C2) integrating the peak partial information s'1(t) and an approximation part s2a(t), a denoised chromatographic peak s' (t) is obtained.
2. The method of improving the signal-to-noise ratio of LC-MS data according to claim 1, wherein in step (B1), F ═ 0.014-0.015-0.1240.0120.5610.6400.141-0.0280.0210.010, L ═ 10.
3. The method for improving signal-to-noise ratio of LC-MS data of claim 1, wherein in step (B2), the peak portion s1(t) and Low pass FilteringConvolving with L.D to obtain a1；
Peak part s1(t) convolved with a high-pass filter H.D to obtain D1。
4. The method for improving the signal-to-noise ratio of LC-MS data of claim 1, wherein in step (C1), the discretization of the basis functions is represented as:
F＝[0.014 -0.015 -0.124 0.012 0.283 0.321 0.141 -0.028 0.021 0.010]，L＝10。
5. the method for improving the signal-to-noise ratio of LC-MS data of claim 1, wherein in step (A1), the peak position p of chromatographic peak s (t) is detected by:
by a functionThe spectral peaks s (t) are subjected to a continuous wavelet transform for the mother wavelet, and a maximum is detected in the wavelet coefficient matrix, which maximum corresponds to the peak position p of the spectral peak s (t).
6. The method for improving the signal-to-noise ratio of LC-MS data of claim 1, wherein in step (A1), the width w of chromatographic peak s (t) is detected by:
by a functionAnd obtaining a wavelet transform coefficient matrix for the mother wavelet, and obtaining the width w of a chromatographic peak s (t).
Technical Field
The invention relates to chromatography, in particular to a method for improving LC-MS data signal-to-noise ratio.
Background
Liquid chromatography-mass spectrometry (LC-MS) is an important tool for qualitative analysis of complex mixture samples. The device integrates the advantages of two instruments of liquid chromatography and mass spectrometry when used independently, has greater advantages, and is widely applied to various fields of environment, food, geological detection and the like. The raw LC-MS data is the basis for qualitative analysis of LC-MS, but it may contain interference from instrument noise, neutrals, other compounds, etc., and if the raw LC-MS data is directly applied to qualitative analysis of a mixture sample, the resulting analysis results are unreliable. Furthermore, the complexity of LC-MS systems often makes it difficult for analysts to meet method detection limitations. Therefore, there is a need for an algorithm that can improve the signal-to-noise ratio of LC-MS while ensuring that the signal is not distorted.
In order to improve the signal-to-noise ratio of LC-MS data, some denoising algorithms are commonly used for processing, including:
1. fourier filtering methods, which have insufficient processing power for non-stationary signals and cannot obtain the time when each frequency occurs.
2. The wavelet denoising method can process non-stationary signals, but the threshold parameter selection of wavelet transformation is difficult.
3. The polynomial smoothing algorithm, also known as SG filtering, weights the data within the window to smooth out the noise, but it is not adaptive.
In addition, the above algorithms all process the whole spectral line at the same time, and may cause signal distortion while filtering noise.
The signal of the LC-MS data has the characteristic of local mutation, and when the signal is processed by using a traditional denoising algorithm, the noise is removed, and the tip of the signal is usually removed, so that the response of a signal peak is reduced, and the subsequent analysis is not facilitated.
Disclosure of Invention
In order to overcome the defects in the prior art, the invention provides the method for improving the signal-to-noise ratio of the LC-MS data, which has good denoising effect and undistorted useful signals.
The purpose of the invention is realized by the following technical scheme:
the method for improving the signal-to-noise ratio of LC-MS data comprises the following steps:
(A1) detecting the peak position p and the width w of the chromatographic peak s (t) to obtain the starting position w of the chromatographic peak s (t)1And an end position w2；
(A2) Dividing the chromatographic peak s (t) into peak portions s1(t) and non-peak portions s2(t) proceeding to step (B1) and step (C1), respectively;
(B1) designing a basis function according to the chromatographic peak s (t), wherein the discretization of the basis function is represented as F, and four orthogonal filters with the length of L are formed according to F; L-D, H-D and L-R, H-R, where L-R is a normalized representation of F, H-R is an orthogonal inverse filter of L-R, L-D is the inverse of L-R, and H-D is the inverse of H-R;
(B2) obtaining a peak portion s1First layer approximation coefficient a of (t)1And first layer detail coefficient d1；
L is the filter length;
(B3) obtaining the peak portion s by means of the step (B2)1(t) multilayer approximation coefficient anAnd each layer detail coefficient d1，d2···dnN is an integer greater than 2;
(B4) will approximate the coefficient anAnd each layer detail coefficient d1，d2···dnRespectively reconstructing to obtain peak portions s1(t) approximate fraction s1a(t) and details s1b(t)；
Detail section s1b(t) performing EMD empirical mode decomposition,imf is the modal component at different scales, rn(t) is the residual component;
(B5) calculating an energy value for each modal componentObtaining the energy mutation occurring at the k imf component;
(B6) reconstructing low-frequency imf components behind the kth imf component and residual error items thereof to obtain detail part information after effective information is extractedThereby obtaining the denoised peak partial information s'1(t)＝s1a(t)+s′1b(t)；
(C1) For non-peak portions s2(t) obtaining non-peak portions s in the manner of steps (B1) - (B4)2(t) approximate fraction s2a(t) and the details;
(C2) integrating the peak partial information s'1(t) and an approximation part s2a(t), a denoised chromatographic peak s' (t) is obtained.
Compared with the prior art, the invention has the beneficial effects that:
the denoising effect is good, and useful signals are not distorted;
the method comprises the steps of identifying the position of a signal peak by adopting a peak detection algorithm, constructing different basis functions for a peak part and a non-peak part according to the characteristics of the signal peak for processing, adopting a multi-level denoising algorithm for extracting useful information for signal waves, and adopting forced denoising processing for the non-wave position, so that the noise is effectively eliminated, the signal distortion is not caused, and the signal-to-noise ratio is effectively improved;
for the processing of signal peaks, imf component decomposition and reconstruction are carried out on the detail part of the scale where the noise exists, so that the difficulty in selecting a threshold coefficient can be avoided, and meanwhile, the useful signal of the scale where the noise exists is reserved, so that the signal is not distorted.
Drawings
The disclosure of the present invention will become more readily understood with reference to the accompanying drawings. As is readily understood by those skilled in the art: these drawings are only for illustrating the technical solutions of the present invention and are not intended to limit the scope of the present invention. In the figure:
FIG. 1 is a diagram of raw signals according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a signal after wavelet transform;
fig. 3 is a schematic diagram of signals processed by a method according to an embodiment of the invention.
Detailed Description
Fig. 1-3 and the following description depict alternative embodiments of the invention to teach those skilled in the art how to make and use the invention. Some conventional aspects have been simplified or omitted for the purpose of teaching the present invention. Those skilled in the art will appreciate that variations or substitutions from these embodiments will be within the scope of the invention. Those skilled in the art will appreciate that the features described below can be combined in various ways to form multiple variations of the invention. Thus, the present invention is not limited to the following alternative embodiments, but is only limited by the claims and their equivalents.
Example 1:
the structure diagram of the method for improving the signal-to-noise ratio of LC-MS data of the embodiment of the invention comprises the following steps:
(A1) detecting the peak position p and the width w of the chromatographic peak s (t) to obtain the starting position w of the chromatographic peak s (t)1And an end position w2；
(A2) Dividing the chromatographic peak s (t) into peaksSection s1(t) and non-peak portions s2(t) proceeding to step (B1) and step (C1), respectively;
(B1) designing a basis function according to the chromatographic peak s (t), wherein the discretization of the basis function is represented as F, and four orthogonal filters with the length of L are formed according to F; L-D, H-D and L-R, H-R, where L-R is a normalized representation of F, H-R is an orthogonal inverse filter of L-R, L-D is the inverse of L-R, and H-D is the inverse of H-R;
(B2) obtaining a peak portion s1First layer approximation coefficient a of (t)1And first layer detail coefficient d1；
L is the filter length;
(B3) obtaining the peak portion s by means of the step (B2)1(t) multilayer approximation coefficient anAnd each layer detail coefficient d1，d2···dnN is an integer greater than 2;
(B4) will approximate the coefficient anAnd each layer detail coefficient d1，d2···dnRespectively reconstructing to obtain peak portions s1(t) approximate fraction s1a(t) and details s1b(t)；
Detail section s1b(t) performing EMD empirical mode decomposition,imf is the modal component at different scales, rn(t) is the residual component;
(B5) calculating an energy value for each modal componentObtaining the energy mutation occurring at the k imf component;
(B6) reconstructing low-frequency imf components behind the kth imf component and residual error items thereof to obtain detail part information after effective information is extractedThereby obtaining the denoised peak partial information s'1(t)＝s1a(t)+s′1b(t)；
(C1) For non-peak portions s2(t) obtaining non-peak portions s in the manner of steps (B1) - (B4)2(t) approximate fraction s2a(t) and the details;
(C2) integrating the peak partial information s'1(t) and an approximation part s2a(t), a denoised chromatographic peak s' (t) is obtained.
In order to reduce noise without distortion, further, in step (B1),
F＝[0.014 -0.015 -0.124 0.012 0.561 0.640 0.141 -0.028 0.021 0.010]，L＝10。
in order to reduce noise without distortion, further, in step (B2), the peak portion s1(t) convolved with a low-pass filter L.D to obtain a1；
Peak part s1(t) convolved with a high-pass filter H.D to obtain D1。
In order to reduce noise without distortion, further, in step (C1), the discretization of the basis function is expressed as:
F＝[0.014 -0.015 -0.124 0.012 0.283 0.321 0.141 -0.028 0.021 0.010]，L＝10。
for accurate peak finding, further, in step (a1), the peak position p of the chromatographic peak s (t) is detected by:
by a functionPerforming continuous wavelet transform on chromatographic peak s (t) for mother wavelet, and detecting maximum value in wavelet coefficient matrix, wherein the maximum value corresponds to peak position p of chromatographic peak s (t)
In order to accurately obtain the width of the peak, further, in step (a1), the width w of the chromatographic peak s (t) is detected by:
by a functionAnd obtaining a wavelet transform coefficient matrix for the mother wavelet, and obtaining the width w of a chromatographic peak s (t).
Example 2:
the application of the method for improving the signal-to-noise ratio of LC-MS data in pesticide residue detection in embodiment 1 of the invention is provided.
In the application example, the method for improving the signal-to-noise ratio of LC-MS data comprises the following steps:
(A1) obtaining a chromatographic peak s (t), detecting the peak position p and the width w of the chromatographic peak s (t) as shown in FIG. 1, and obtaining the starting position w of the chromatographic peak s (t)1And an end position w2；
The peak position p is obtained in the following manner:
by a functionPerforming continuous wavelet transform on the chromatographic peak s (t) for the mother wavelet, and detecting a maximum value in a wavelet coefficient matrix, wherein the maximum value corresponds to the peak position p of the chromatographic peak s (t);
the width w of the chromatographic peak s (t) is obtained in the following manner:
by a functionObtaining a wavelet transform coefficient matrix for the mother wavelet to obtain the width w of a chromatographic peak s (t);
(A2) dividing the chromatographic peak s (t) into peak portions s1(t) and non-peak portions s2(t) proceeding to step (B1) and step (C1), respectively;
(B1) designing a basis function according to the chromatographic peak s (t), wherein the discretization of the basis function is expressed as F ═ 0.014-0.015-0.1240.0120.5610.6400.141-0.0280.0210.010], and four orthogonal filters with the length L are formed according to F; L-D, H-D and L-R, H-R, where L-R is a normalized representation of F, H-R is an orthogonal inverse filter of L-R, L-D is the inverse of L-R, and H-D is the inverse of H-R; l ═ 10;
(B2) peak part s1(t) convolving with a low-pass filter L.D to obtain a peak portion s1First layer approximation coefficient a of (t)1；L is the filter length;
peak part s1(t) convolving with a high-pass filter H.D to obtain a first layer detail coefficient D1，L is the filter length;
(B3) obtaining the peak portion s by means of the step (B2)1(t) multilayer approximation coefficient anAnd each layer detail coefficient d1，d2···dnN is an integer greater than 2; the embodiment is three layers;
(B4) will approximate the coefficient anAnd each layer detail coefficient d1，d2···dnRespectively reconstructing to obtain peak portions s1(t) approximate fraction s1a(t) and details s1b(t)；
Detail section s1b(t) performing EMD empirical mode decomposition,imf is the modal component at different scales, rn(t) is the residual component;
(B5) calculating an energy value for each modal componentObtaining the energy mutation occurring at the k imf component;
(B6) reconstructing low-frequency imf components behind the kth imf component and residual error items thereof to obtain detail part information after effective information is extractedThereby obtaining the denoised peak partial information s'1(t)＝s1a(t)+s′1b(t)；
(C1) For non-peak portions s2(t) obtaining non-peak portions s in the manner of steps (B1) - (B4)2(t) approximate fraction s2a(t) And a detailed section;
for non-peak portions s2Discretization of the basis functions of (t);
F＝[0.014 -0.015 -0.124 0.012 0.283 0.321 0.141 -0.028 0.021 0.010]，L＝10。
(C2) integrating the peak partial information s'1(t) and an approximation part s2a(t), a denoised chromatographic peak s' (t) is obtained, as shown in FIG. 3.
Fig. 2 schematically shows a signal after processing an original signal using wavelet transformation, where the signal-to-noise ratio is low and the signal is distorted in non-peak portions as compared with the processing of the present embodiment.
- 上一篇：一种医用注射器针头装配设备
- 下一篇：一种盐酸贝尼地平片中的杂质检测方法