WO2020199866A1 - Biometabolomics data processing and analysis methods and apparatuses, and application thereof - Google Patents

Abstract

Biometabolomics data processing and analysis methods and apparatuses, and an application thereof. The biometabolomics data processing method comprises a step of integrating liquid chromatography-mass spectrometry data or gas chromatography-mass spectrometry data of a plurality of biological samples to form a feature database. The integration step comprises: S11, randomly selecting one of the plurality of biological samples as a reference sample, and correcting the timelines of other samples one by one according to the timeline of the reference sample; S12, performing peak recognition processing of a primary mass spectrometry ion peak on the corrected samples one by one to obtain a plurality of recognition feature peaks; and S13, merging the plurality of recognition feature peaks according to the sample information complementation principle to obtain a feature database of the plurality of biological samples. The processing method can realize integration of mega-scale metabolomics data, and can achieve data correction and data integration in batches or of a single sample, without being affected by detection batches.

Description

Biological metabolomics data processing method, analysis method, device and application Technical field

The present invention relates to the technical field of metabolomics, in particular to a method, analysis method, device and application of biological metabolomics data processing.

Background technique

Metabolomics is a new subject after genomics and proteomics. It is an important part of systems biology. It mainly investigates the dynamic changes of all small molecule metabolites and their contents before and after the biological system is stimulated or disturbed. Through the overall qualitative and quantitative analysis of all small molecule metabolites in the organism, the relationship between metabolites and physiological and pathological changes can be explored and discovered. Studies have shown that metabolome has important application value in the fields of early disease diagnosis, biomarker discovery, drug screening, toxicity evaluation, sports medicine, and nutrition.

With the rapid development of science and technology, research and detection methods for metabolome emerge in endlessly. At present, the most widely used and most powerful is mainly liquid chromatography-mass spectrometry (LC-MS). In recent years, LC-MS technology has been further improved, and large-scale sample detection applications have also increased. As the number of test samples increases, a series of problems have also arisen. For example, the test time for large-scale samples is longer, and the sensitivity of the machine will decrease and retention time drift during long-term operation. Therefore, researchers often put large-scale samples on the machine in batches to keep the machine running in good condition, but this has another problem, that is, the metabolome data between samples and batches are random. Errors and systematic errors cannot be directly compared, and data integration is required. For the integration of data between different samples and different batches, there are currently some methods that can be used. The common one is to use the XCMS method for data integration, which can realize multi-sample metabolomics data analysis.

However, using methods such as XCMS to integrate metabolome data from different samples and different batches also has some problems and limitations. Their current processing method is to put all the sample data together for integration, and cannot be integrated separately in batches or individual samples. For a fixed number of samples, it can be processed, and the processing time varies depending on the size of the sample number. The disadvantage of this processing method is that the data processing time and difficulty will increase with the increase of the number of samples. When the number of samples is very large or there are constantly new samples that need data integration, this method may not be suitable, and Not conducive to commercial applications. At the same time, the existing methods still have some problems and shortcomings. For example, they cannot effectively use the sample information complementation between different batches. Different batches of samples have their own coordinates. The information is difficult to compare and complementary, and some information will be lost. As a result, the repeatability and coverage of metabolite detection will be reduced.

In order to solve the above problems, the present invention provides a biological metabolomics data processing method, analysis method and device, which can effectively solve the problem that the sample information complementation between different batches cannot be effectively used in the metabolome data processing process, resulting in poor metabolite detection repeatability And the coverage will be reduced.

Summary of the invention

The present invention aims to provide a biological metabolomics data processing method, analysis method, device and application, which are suitable for processing larger scale metabolomics data.

In order to achieve the above objective, according to one aspect of the present invention, a method for processing biological metabolomics data is provided. The biological metabolomics data includes liquid chromatography-mass spectrometry data or gas chromatography-mass spectrometry data, liquid chromatography-mass spectrometry data includes primary mass spectrometry data, and gas chromatography-mass spectrometry data includes primary mass spectrometry data; biological metabolomics data processing method It includes the steps of integrating liquid chromatography-mass spectrometry data or gas chromatography-mass spectrometry data of multiple biological samples to form a feature database. The integration steps include:

S11: arbitrarily select one of the multiple biological samples as a reference sample, and perform correction on the time axis of other samples one by one according to the time axis of the reference sample;

S12, for each sample after calibration, perform peak identification processing of the ion peaks of the primary mass spectrum one by one to obtain multiple identification characteristic peaks; and

S13: According to the principle of complementarity of sample information, a plurality of identification characteristic peaks are combined to obtain a characteristic database of a plurality of biological samples.

Further, in S13: if the [mzmin, mzmax] regions of the multiple identification feature peaks overlap or are adjacent, and the [rtmin, rtmax] regions overlap or are adjacent, then the multiple identification feature peaks are merged into one feature peak.

Further, S13 includes:

S131: Determine whether the [mzmin, mzmax] regions of multiple identification characteristic peaks overlap or are adjacent, if they overlap, go to S133; if they do not overlap, further determine whether they are adjacent, if multiple [mzmin, mzmax] regions of multiple identification characteristic peaks If the interval between is less than the first preset threshold, it is determined to be adjacent and enter S133; if it is neither overlapping nor adjacent, it is determined that the multiple identification characteristic peaks are independent characteristic peaks;

S132: Determine whether the [rtmin, rtmax] regions of the multiple identification feature peaks overlap or are adjacent, if they overlap, go to S133; if they do not overlap, further determine whether they are adjacent, if the [rtmin, rtmax] regions of the multiple identification feature peaks If the interval between is less than the second preset threshold, it is judged to be adjacent, and enter S133; if it is neither overlapping nor adjacent, it is judged that the multiple identification characteristic peaks are independent characteristic peaks;

S133: If the multiple identification characteristic peaks simultaneously satisfy the overlapping or adjacent conditions in S131 and the overlapping or adjacent conditions in S132, the multiple identification characteristic peaks are merged into one characteristic peak;

S134: Generate a feature list using data of all feature peaks to obtain a feature database.

Further, the first preset threshold is set according to instrument parameters, and the second preset threshold is set according to the maximum value of the time deviation in the retention time correction; preferably, the first preset threshold is set to 0.01～0.015Da , The second preset threshold is set to 10-15.

Further, the mass spectrometry data also includes secondary mass spectrometry data, and S13 also includes:

S135. Compare the secondary mass spectrometry data of the multiple biological samples to the feature database generated in S134, wherein when the comparison rate is greater than or equal to a third preset threshold, it is determined that the peak combination is effective.

Further, the third preset threshold is set to 40%.

Further, the third preset threshold is set to 50%.

Further, the third preset threshold is set to 60%.

Further, the third preset threshold is set to 80%.

Further, the mass spectrum data also includes the secondary mass spectrum data, and S11 also includes the retention time correction of the primary mass spectrum data and the secondary mass spectrum data; preferably, the retention time correction is performed using the Obiwarp algorithm.

Further, the algorithm for peak recognition is CentWave algorithm, matchedFilter algorithm or mzMine algorithm.

Further, the parameter settings of the peak recognition algorithm include: ppm: the resolution of the instrument used; peak width: set to 2-30; noise: set to 0; signal-to-noise ratio: set to 10.

Further, biological samples include human or animal body fluids, tissues or cells, plant roots, stems, leaves, fruits or seeds, or microbial cell culture fluid; wherein, body fluids include urine, blood, saliva, cerebrospinal fluid or amniotic fluid, Tissues include organ tissues, muscle tissues or tumor tissues, and cells include stem cells, somatic cells, tumor cells or microbial cells.

According to another aspect of the present invention, a method for analyzing biological metabolomics data is provided. The analysis method sequentially includes the steps of biological metabolomics data processing and qualitative identification of metabolites through secondary mass spectrometry data information, wherein the biological metabolomics data processing adopts any of the above-mentioned biological metabolomics data processing methods of the present invention. .

Further, the step of qualitatively identifying metabolites through the data information of the secondary mass spectrum includes:

S21, obtain the mass-to-charge ratio data of each standard compound;

S22. Select a characteristic value arbitrarily in the characteristic database obtained after the biological metabolomics data processing, and find all the mass-to-charge ratio data of the second-stage mass spectrometry corresponding to the characteristic value, and according to all the mass-to-charge ratio data of the second-stage mass spectrum, Find a matching standard compound;

S23, take all the mass-to-charge ratio data of the MS mass spectrum corresponding to a characteristic value selected in S22 as one side, and use the mass-to-charge ratio data of the MS mass spectrum of the matched standard compound found in S22 as the other side, and perform Similarity scores, points points are calculated, and metabolites are qualitatively based on the points.

Further, S23 includes: calculating the median of the similarity between each standard compound in the multiple standard compounds on the matching and multiple MS mass spectrometry data, and selecting the compound with the largest median; preferably, according to the median of the compound Whether it is greater than the cut-off value, judge whether it matches.

Furthermore, the mass-to-charge ratio data of standard compounds are obtained from existing databases, including NISTlib, HMDB or METLIN.

Further, the analysis method also includes a step of quantifying biological metabolites.

Further, the steps of quantifying biological metabolites include:

S31: Correct the time axis of the sample to be quantified according to the time axis of the reference sample;

S32: Integrating the corresponding characteristic regions of the sample to be quantified in the established characteristic database to obtain a relatively quantitative result of biological metabolites.

According to another aspect of the present invention, there is provided a method for processing data of the above-mentioned biological metabolomics and analysis method for data of biological metabolomics in terms of vitamins, amino acids, lipids, steroids, aromatic acids, neurotransmitters, pigments, and carbohydrates. Or the application of short peptide identification.

According to another aspect of the present invention, a method for detecting vitamins, amino acids, lipids, steroids, aromatic acids, neurotransmitters, pigments, carbohydrates or short peptides is provided. The detection method includes: performing liquid chromatography-mass spectrometry and/or gas chromatography-mass spectrometry on a biological sample to obtain liquid chromatography-mass spectrometry data and/or gas chromatography-mass spectrometry data; using any of the above-mentioned biological metabolomics data processing Method or analysis method of biological metabolomics data process the liquid chromatography-mass spectrometry data and/or gas chromatography-mass spectrometry data of biological samples to obtain data results; and convert vitamins, amino acids, lipids, steroids, and aromas based on the data results Acid, neurotransmitter, pigment, carbohydrate or short peptide.

According to another aspect of the present invention, a biological metabolomics data processing device is provided. Among them, biological metabolomics data includes liquid chromatography-mass spectrometry data or gas chromatography-mass spectrometry data, liquid chromatography-mass spectrometry data includes primary mass spectrometry data, and gas chromatography-mass spectrometry data includes primary mass spectrometry data; biological metabolomics data processing The device includes a database generating module that integrates liquid chromatography-mass spectrometry data or gas chromatography-mass spectrometry data of multiple biological samples to form a feature database. The database generating module includes:

The time axis correction sub-module is set to arbitrarily select one of multiple biological samples as a reference sample, and correct the time axis of other samples one by one according to the time axis of the reference sample;

The characteristic peak recognition sub-module is configured to perform peak recognition processing of the ion peaks of the primary mass spectrometer one by one for each sample after calibration to obtain multiple characteristic peaks; and

The feature database forming sub-module is set to merge multiple identification feature peaks according to the principle of sample information complementarity to obtain feature databases of multiple biological samples.

Further, the feature database forming sub-module includes a data integration unit, the data integration unit is set to overlap or adjacent [mzmin, mzmax] regions, and [rtmin, rtmax] regions overlap or merge multiple adjacent identification feature peaks into one Characteristic peaks.

Further, the feature database forming sub-module includes a first judgment unit, a second judgment unit, a data integration unit and a feature database forming unit:

Among them, the first judging unit is set to judge whether the [mzmin, mzmax] regions of multiple identification feature peaks overlap, if they overlap, enter the data integration unit; if they do not overlap, further judge whether they are adjacent, if multiple identification feature peaks are If the interval of the [mzmin, mzmax] area is less than the first preset threshold, it is determined to be adjacent and enter S133; if it is neither overlapping nor adjacent, it is determined that the multiple identification characteristic peaks are independent characteristic peaks;

The second judging unit is set to judge whether the [rtmin, rtmax] regions of multiple identification feature peaks overlap or are adjacent, if they overlap, enter the data integration unit; if they do not overlap, further judge whether they are adjacent, if multiple identification feature peaks If the interval of the [rtmin, rtmax] area is less than the second preset threshold, it is judged to be adjacent and enter the data integration unit; if it is neither overlapping nor adjacent, it is judged that the multiple identification characteristic peaks are independent characteristic peaks;

The data integration unit is set to combine the [mzmin, mzmax] area overlap or adjacent, and the [rtmin, rtmax] area overlap or the adjacent multiple identification characteristic peaks are combined into one characteristic peak;

The feature database forming unit is configured to generate a feature list using the data of all feature peaks to obtain the feature database.

Further, the first preset threshold is set according to instrument parameters, and the second preset threshold is set according to the maximum value of the time deviation in the retention time correction; preferably, the first preset threshold is set to 0.01～0.015Da , The second preset threshold is set to 10-15.

Further, the mass spectrometry data also includes secondary mass spectrometry data, and the biological metabolomics data processing device further includes: a peak merging validity verification sub-module configured to compare the secondary mass spectrometry data of multiple biological samples to a feature database, wherein , When the comparison rate is greater than or equal to the third preset threshold, it is determined that the peak combination is valid.

Further, the third preset threshold is set to 40%.

Further, the third preset threshold is set to 50%.

Further, the third preset threshold is set to 60%.

Further, the third preset threshold is set to 80%. Further, the mass spectrum data also includes secondary mass spectrum data, and the time axis correction sub-module is also configured to perform retention time correction on the primary mass spectrum data and the secondary mass spectrum data; preferably, the Obiwarp algorithm is used to perform retention time correction.

Further, the algorithm for peak recognition is CentWave algorithm, matchedFilter algorithm or mzMine algorithm.

Further, the parameter settings of the peak recognition algorithm include: ppm: the resolution of the instrument used; peak width: set to 2-30; noise: set to 0; signal-to-noise ratio: set to 10.

According to another aspect of the present invention, a device for analyzing biological metabolomics data is provided. The analysis device includes a module configured to process biological metabolomics data and a module configured to qualitatively identify metabolites through secondary mass spectrometry data information, wherein the module configured to process biological metabolomics data is any of the above-mentioned biological Metabolomics data processing device.

Further, the module configured to qualitatively identify metabolites through the data information of the secondary mass spectrum includes:

The standard compound mass-to-charge ratio data acquisition sub-module is set to acquire the mass-to-charge ratio data of each standard compound;

The standard compound matching sub-module is set to randomly select a characteristic value in the characteristic database obtained after the biological metabolomics data processing, and find all the MS mass-to-charge ratio data corresponding to the characteristic value, according to all the secondary Mass-to-charge ratio data of mass spectrometry to find a standard compound that matches it;

The integral qualitative sub-module is set to take all the MS mass-to-charge ratio data corresponding to a characteristic value selected in the standard compound matching sub-module as one side, and use the standard compound matching sub-module to find the second level of the matched standard compound The mass-to-nucleus ratio data of the mass spectrometer is the other party, score the similarity between the two, calculate the point integration, and qualitative the metabolites based on the integrated value.

Further, the integral qualitative sub-module is set to calculate the median of the similarity between each standard compound and the multiple secondary mass spectrometry data among the multiple standard compounds on the matching, and select the compound with the largest median; preferably, according to the compound Whether the median is greater than the cut-off value, judge whether it matches.

Furthermore, the mass-to-charge ratio data of standard compounds are obtained from existing databases, including NISTlib, HMDB or METLIN.

Further, the analysis device further includes a module configured to quantify biological metabolites.

Further, the module set to quantify biological metabolites includes:

The time axis correction sub-module is set to correct the time axis of the sample to be quantified according to the time axis of the reference sample;

The relative quantification of biological metabolites sub-module is set to integrate the corresponding characteristic regions of the samples to be quantified in the established characteristic database to obtain the relative quantitative results of biological metabolites.

By implementing the above technical methods, the present invention has at least the following beneficial effects:

Applying the technical solution of the present invention, by constructing a feature database, unifying the time axis, using the principle of information complementarity between samples for peak merging, etc., ultra-large-scale metabolome data integration can be realized, and batches can be realized Or the data correction and data integration of a single sample is not affected by the test batch, and is suitable for commercial testing.

The invention constructs a feature database, fixes a reference sample, and unifies the time axis, which can ensure that subsequent samples are comparable in time, so that the metabolome data processing process can effectively use sample information complementation between different batches, and effectively improve metabolism. Object detection repeatability and coverage.

The present invention performs peak merging processing in the process of constructing a feature database, and the merged peaks can cover a larger area, so that quantification can be performed more accurately when only one sample is detected, even if the chromatographic peak shape is not good Metabolites still have a good effect, and produce a larger coverage area, which makes it more compatible with subsequent samples, and effectively reduces the impact of retention time (RT) drift.

After the feature database is established, the present invention effectively improves the analysis efficiency of samples, so that subsequent samples are comparable in time, and the samples do not need to be rolled back, and can be widely used in business.

Description of the drawings

The accompanying drawings constituting a part of the present application are used to provide a further understanding of the present invention. The exemplary embodiments and descriptions of the present invention are used to explain the present invention, and do not constitute an improper limitation of the present invention. In the attached picture:

Fig. 1 shows a schematic diagram of the process of constructing a feature database in an embodiment of the present invention;

FIG. 2 shows a schematic diagram of the process of merging and identifying characteristic peaks in an embodiment of the present invention;

Figure 3 shows a sample retention time correction diagram in Embodiment 1;

Figure 4 shows the ionization form diagrams of the 16 standard compounds in Example 1;

Figure 5 shows the similarity comparison diagram between 35 MS2 of Example 1 and 16 standard compounds on matching;

FIG. 6 shows the distribution of the number of feature missing values in Embodiment 1;

Figure 7 shows a comparison graph of the coefficient of variation (CV) between the samples of Example 1 and Comparative Example 1;

Figure 8 shows the PCA results of Example 1 and Comparative Example 1;

Figure 9 shows a comparison of the number of metabolites identified in Example 1 and Comparative Example 1;

Figure 10 shows the mz and RT distributions of MS2 precursor ions corresponding to FT08341 in Example 1; and

Figure 11 shows the similarity of 35 MS2 spectra of FT08341 in Example 1.

detailed description

It should be noted that the embodiments in this application and the features in the embodiments can be combined with each other if there is no conflict. Hereinafter, the present invention will be described in detail with reference to the drawings and in conjunction with the embodiments.

The abbreviations and terms involved in the present invention are explained as follows:

Metabolome: refers to the dynamic overall of metabolites in an organism. The metabolome usually refers to the general term for small molecular metabolites with a relative molecular mass of less than 1000 Da (Da: Dalton).

Mass Spectrometry: also known as Mass Spectrometry (MS), is a method of ionizing the measured substance, using electric and magnetic fields to separate the moving ions according to their mass-to-charge ratios for detection.

Precursor ion: also known as precursor ion, is an ion that can further decompose and generate fragment ions.

Product ion: Fragment ion obtained by high-energy fragmentation of a certain molecular ion (parent ion).

Primary mass spectrum (MS1): Detect the mass-to-charge ratio and intensity of all charged ions to form a primary spectrum. The signal in the primary mass spectrum is the precursor ion signal.

Second-level mass spectrometry (MS2): Select parent ions in a certain way, dissociate them further, analyze the mass-to-charge ratio and intensity of the formed product ions, and form a second-level spectrum.

Mass-to-charge ratio (mz): The ratio of the mass of a charged ion to the charge, which is the physical characteristic of the ion and is a certain value. Limited by the resolution of the instrument, the detected mz will fluctuate.

Retention Time (Retention Time, RT): the time from the beginning of the sample injection to the time when the maximum concentration of the component appears after the column, that is, from the beginning of the sample injection to the peak of a certain component chromatographic peak The elapsed time. For a specific separation column, the retention time of the component (molecular ion) is related to its physical and chemical properties.

Ion peaks: Ion peaks in a sample, expressed in [mzmin, mzmax, rtmin, rtmax].

Features (features): have the same representation form as peaks [mzmin, mzmax, rtmin, rtmax]. Unlike peaks, features can represent the molecular ion (peaks is a part of the molecular ion, and a molecular ion can have multiple peaks). Features can be merged from multiple peaks from one sample, or from multiple peaks from multiple samples.

PPM: parts per million, is the expression of the ratio, which means "parts per million...".

Based on the LC-MS technology, the current metabolome testing of large-scale samples is performed sequentially and in batches. There are deviations between samples and batches, and the sample data of the same batch and different batches need to be integrated Only then can the next step of comparative analysis be carried out. For the integration and analysis of large-scale metabolomics data, some existing technologies (such as XCMS) have drawbacks. They need to put all sample data together for integration. They cannot be integrated in batches or individual samples individually. There are also some shortcomings in the qualitative and quantitative aspects of metabolites.

In view of these shortcomings in the prior art, the present invention proposes a new metabolome data integration idea, which can be applied to large-scale metabolome data analysis, and can realize data correction and data integration in batches or single samples, and It is not affected by the test batch, and at the same time, the coverage of metabolites and the accuracy of qualitative and quantitative are also improved.

According to a typical embodiment of the present invention, a method for processing biological metabolomics data is provided. Biological metabolomics data includes liquid chromatography-mass spectrometry data and/or gas chromatography-mass spectrometry data, liquid chromatography-mass spectrometry data includes primary mass spectrometry data, and gas chromatography-mass spectrometry data includes primary mass spectrometry data; the biological metabolomics data The processing method includes the step of integrating liquid chromatography-mass spectrometry data or gas chromatography-mass spectrometry data of multiple biological samples to form a feature database, and the integration step includes:

S11: arbitrarily select one of the multiple biological samples as a reference sample, and perform correction on the time axis of other samples one by one according to the time axis of the reference sample;

S12, for each sample after calibration, perform peak identification processing of the ion peak of the first-level mass spectrum one by one to obtain multiple identification characteristic peaks;

And S13, according to the principle of complementary sample information, the multiple identification characteristic peaks are combined to obtain a characteristic database of multiple biological samples.

Applying the technical solution of the present invention, first select a sample as a reference sample, and the time axis of other samples are corrected according to this sample, that is, a unified coordinate axis is determined so that the liquid chromatography-mass spectrometry data or gas chromatography-mass spectrometry data of the sample is Comparable in time; on the corrected time axis, peak identification is performed on the primary mass spectrum ion peak of each sample, and then peaks are merged using the principle of information complementarity between samples to construct a feature database, thereby It can realize the integration of super large-scale metabolome data. Since all samples are calibrated on the time axis according to the reference sample, data calibration and data integration can be realized in batches or single samples, and are not affected by the testing batch, which is suitable for commercial testing.

In an embodiment of the present invention, the mass spectrometry data also includes secondary mass spectrometry data, and S11 also includes retention time correction of the primary and secondary mass spectrum data to further improve the accuracy of the mass spectrometry data; preferably, the Obiwarp algorithm is used for Retention time correction has the advantages of fast calculation speed and high data processing accuracy.

In an embodiment of the present invention, S13 includes: if the [mzmin, mzmax] regions of the multiple identification feature peaks overlap or are adjacent, and the [rtmin, rtmax] regions overlap or are adjacent, then merge the multiple identification feature peaks into A characteristic peak; preferably, S13 includes: S131, judging whether the [mzmin, mzmax] regions of multiple identifying characteristic peaks overlap, if they overlap, go to S133; if they do not overlap, further determine whether they are adjacent, if the multiple identified If the interval between the [mzmin, mzmax] regions of the characteristic peaks is less than the first preset threshold, it is determined to be adjacent, and then proceeds to S133; if it is neither overlapping nor adjacent, it is determined that multiple identification characteristic peaks are independent characteristic peaks; S132: Determine whether the [rtmin, rtmax] regions of the multiple identification feature peaks overlap or are adjacent, if they overlap, go to S133; if they do not overlap, further determine whether they are adjacent, if the [rtmin, rtmax] regions of the multiple identification feature peaks If the interval is less than the second preset threshold, it is judged to be adjacent and enter S133; if it is neither overlapping nor adjacent, it is judged that the multiple identification feature peaks are independent feature peaks; S133, if multiple identification feature peaks are At the same time, if the two conditions of overlap or adjacent in S131 and overlap or adjacent in S132 are met, multiple identification characteristic peaks are merged into one characteristic peak; S134, the characteristic database is obtained by generating a characteristic list using the data of all characteristic peaks. In this way, the combined characteristic peaks can cover a larger area, even if only one sample is detected, it can be quantified more accurately (effective for metabolites with poor chromatographic peak shapes), and a larger coverage area can be more effective Compatible with subsequent samples, effectively reducing the impact of retention time (RT) drift.

Preferably, the peak recognition algorithm is CentWave algorithm, matchedFilter algorithm or mzMine algorithm, more preferably CentWave algorithm, because this method can improve sensitivity, limit its error, and find the most recognized characteristic peaks most accurately. In this way, peaks are merged on the basis of CentWave algorithm, which can effectively use CentWave to locate the maximum response area. In the present invention, the algorithm parameter setting follows the idea of "improving detection sensitivity as much as possible". Preferably, the parameter setting of the peak recognition algorithm includes: ppm: the resolution of the instrument is adopted; peak width: set to 2-30; noise: Set to 0; SNR: set to 10.

In an embodiment of the present invention, the first preset threshold is set according to instrument parameters, and the second preset threshold is set according to the maximum value of the time deviation in the retention time correction; preferably, the first preset threshold is set It is 0.01~0.015Da, more preferably, the first preset threshold is set to 0.01Da, 0.011Da, 0.012Da, 0.013Da, 0.014Da or 0.015Da, and the second preset threshold is set to 10~15, more preferably Yes, the second preset threshold is set to 10, 11, 12, 13, 14 or 15, to improve the effectiveness of peak merging, thereby improving the accuracy of the feature database.

In an embodiment of the present invention, preferably, the mass spectrometry data further includes secondary mass spectrometry data, and S13 further includes: S135, comparing the secondary mass spectrometry data of multiple biological samples to the feature database generated in S134, and assisting in determining peak merging The higher the ratio of the secondary mass spectrum data to the feature database, the stronger the effectiveness of peak merging. In an embodiment of the present invention, the secondary mass spectrometry data of multiple biological samples are compared to the feature database generated in S134, wherein when the comparison rate is greater than or equal to the third preset threshold, it is determined that the peak combination is effective; one is more preferred In the embodiment, the third preset threshold is set to 40%; in a more preferred embodiment, the third preset threshold is set to 50%; in a more preferred embodiment, the third preset threshold is set In a more preferred embodiment, the third preset threshold is set to 80%.

The biological metabolomics data processing method of the present invention is suitable for almost all biological samples that can be detected by liquid chromatography-mass spectrometry and/or gas chromatography-mass spectrometry. These biological samples include but are not limited to human or animal body fluids, tissues or cells , Plant roots, stems, leaves, fruits or seeds, or microbial cell culture fluid, etc.; among them, body fluids include urine, blood, saliva, cerebrospinal fluid or amniotic fluid, etc., tissues include organ tissue, muscle tissue or tumor tissue, etc., cell Including stem cells, somatic cells, tumor cells or microbial cells, etc.

Under the purpose of the invention, a method for analyzing biological metabolomics data is also provided. The biological metabolomics data analysis method sequentially includes the steps of biological metabolomics data processing and qualitative identification of metabolites through secondary mass spectrometry data information, wherein the biological metabolomics data processing adopts any of the above-mentioned biological metabolisms of the present invention. The omics data processing method is carried out. Since the above-mentioned biological metabolomics data processing method of the present invention is not affected by the detection batch, the sample data volume of the characteristic database can be continuously accumulated, thereby continuously improving the qualitative identification of metabolites through the secondary mass spectrum data information. accuracy.

According to an exemplary embodiment of the present invention, the step of qualitatively identifying metabolites through the data information of the secondary mass spectrum includes:

S21, obtain the mass-to-charge ratio data of each standard compound;

S22. Select a characteristic value arbitrarily in the characteristic database obtained after the biological metabolomics data processing, and find all the mass-to-charge ratio data of the second-stage mass spectrometry corresponding to the characteristic value, and according to all the mass-to-charge ratio data of the second-stage mass spectrum, Find a set of matching standard compounds;

S23, take all the mass-to-charge ratio data of the MS mass spectrum corresponding to a characteristic value selected in S22 as one side, and use the mass-to-charge ratio data of the MS mass spectrum of the matched standard compound found in S22 as the other side, and perform Similarity scores, points points are calculated, and metabolites are qualitatively based on the points.

This method can effectively avoid the problem that the median is not representative, and the operation is simple.

In an embodiment of the present invention, a general method of calculating dot product can also be used to score MS2 similarity. This method is subordinated to the comparison of multiple MS2 belonging to the same feature with the MS2 of the standard compound, and the comparison can be achieved through integration. The purpose of feature identification.

Preferably, S23 specifically includes: calculating the median of the similarity between each standard compound in the multiple standard compounds and multiple MS mass spectrometry data, and selecting the compound with the largest median; more preferably, according to the median of the compound Whether the number of digits is greater than the cut-off value, it is judged whether it matches. Using the above steps, not only includes the "representative" MS2, but also adds various possible MS2 of the compound, which increases the degree of matching with the standard compound.

In the present invention, the mass-to-charge ratio data of the standard compound is obtained from an existing database, for example, the database includes NISTlib, HMDB or METLIN.

In a typical embodiment of the present invention, the analysis method further includes a step of quantifying biological metabolites. After the above-mentioned data processing and qualitative steps, a unified time axis is determined, the retention time is corrected, and a feature database with rich data volume is obtained, so as to maximize the coverage area of the precursor ion (mz) , Can reduce the influence of the fluctuation of mass-to-charge ratio mz and retention time RT, and improve the accuracy of biological metabolite quantification. Preferably, the step of quantifying biological metabolites includes: S31, calibrating the time axis of the sample to be quantified according to the time axis of the reference sample; S32, integrating the corresponding feature area of the sample to be quantified in the established feature database to obtain the biological The result of relative quantification of metabolites.

Based on the above description of the technical solution, in an embodiment or embodiment of the present invention, the specific technical solution is as follows:

1. Construct a feature database to integrate multi-sample metabolome data. See Figure 1 for the process of constructing a feature database.

1) Fix the reference sample and unify the coordinate axis. Specifically: determining a unified coordinate axis to make the liquid chromatography-mass spectrometry data or gas chromatography-mass spectrometry data of the sample comparable in time.

In the entire testing process, the sample is selected as the reference sample (the reference sample is the same as the test sample type, which can be understood as a standard product and only needs to be tested once), and the time axis of other samples are corrected according to this sample, which is a reference. xml, that is, fix a reference sample to ensure that subsequent samples are comparable in time.

2) Retention time correction. Specifically, the new sample is first corrected for retention time (RT). In this step, the Obiwarp algorithm is used to perform retention time (RT) correction on the primary mass spectrum data (MS1) and secondary mass spectrum data (MS2).

3) Peak identification. Specifically: on the corrected time axis, use the CentWave algorithm to identify the peaks of each sample's primary mass spectrum ion peak (findPeaks). Among them, the peak recognition algorithm includes but not limited to CentWave, matchedFilter, mzMine, and CentWave is preferred. This method can improve the sensitivity, limit its error, and find the most identifying characteristic peaks (peak1, peak2,..., peakn) most accurately. The noise caused by high sensitivity and the problem of the same ion peak being divided into two ion peaks caused by the strict ppm setting are handled by sample information complementation.

Among them, the algorithm parameter setting follows the idea of "improving detection sensitivity as much as possible":

①ppm: According to the type of the instrument, the resolution of the instrument is adopted to reduce the error tolerance rate.

②peakwidth (peak width): set to 2～30. This parameter setting is related to the column type and elution time, generally 1/10 of the elution time. The purpose of selecting 2 as the lower limit is to identify very narrow peaks and improve the sensitivity of findPeaks.

③noise (noise): set to 0. This parameter represents the intensity of noise, and the purpose of setting it to 0 is to improve sensitivity. The greater the noise, the lower the sensitivity

④ snthresh (signal to noise ratio): set to 10. This parameter represents the signal-to-noise ratio and uses the default parameters.

4) Merge peaks according to the principle of complementary sample information. Specifically, according to the principle of complementary sample information, the peaks are combined and identified to generate a unified coordinate, that is, a feature database. The processing method is as follows (see Figure 2):

For the identification characteristic peaks peak1, peak2,..., peakn from multiple samples, make the following judgments:

① Determine whether the [mzmin, mzmax] regions of multiple identified characteristic peaks overlap or are adjacent, if they overlap, enter ③; if they do not overlap, further determine whether they are adjacent, if peak m+1, peak m+2,..., peak If the interval of the [mzmin, mzmax] area of m+a is less than the first preset threshold, it is judged to be adjacent, and proceed to ③; if it is neither overlapping nor adjacent, it is judged that multiple identified characteristic peaks are independent characteristic peaks. ；

② Determine whether the [rtmin, rtmax] regions of multiple identified characteristic peaks overlap or are adjacent, if they overlap, enter ③; if they do not overlap, determine whether they are adjacent, if peak m+1, peak m+2,..., peak m The interval of the [rtmin, rtmax] area of +a is less than the second preset threshold absRt, then it is judged to be adjacent, and proceed to ③; if it is neither overlapping nor adjacent, it is judged that multiple identified characteristic peaks are independent characteristic peaks. ；

③If the relationship between peak m+1, peak m+2,..., peak m+a satisfies both the overlap or adjacent conditions in ① and the overlap or adjacent conditions in ②, then it is determined that peak m+1, peak m +2,..., peak m+a belong to the same characteristic peak, and the coordinates of the new characteristic peak are the union of those of a, and the characteristic list is generated to obtain the characteristic database;

Among them, n and a are independently valued in positive integers, m is valued in 0 and positive integers, and m<n.

④ Compare the secondary mass spectrum data of multiple samples to the feature database (rectangular area determined by ["mzmin", "mzmax", "rtmin", "rtmax"]). This step can assist in judging the effectiveness of peak merging.

2. Metabolite identification (qualitative).

Identify metabolites through MS2 data information. The specific methods are as follows:

1) Obtain the mass-to-charge ratio mz of the standard compound (the standard compound is obtained from an existing database, the database is mainly NISTlib, or HMDB, METLIN and other public databases).

2) Find a standard compound with the same mass-to-charge ratio mz of the precursor ion of MS2 (parameter setting: absMz = 0.015Da (Da: Dalton), absMz is set according to the instrument parameters, which can be set to the range of 0.01～0.015Da) .

3) Compare the similarity between MS2 obtained in multiple experiments and multiple standard compounds on the match, score the similarity, and calculate the points.

Combining the results of multiple MS2s, select the most matching standard compound for metabolite identification. This step calculates the median of similarity between each compound and multiple MS2s, and selects the compound with the largest median. According to whether the median of the compound is greater than the specified value (also called the cut-off value, the specified value can be determined as 0.5 to 1 according to the actual situation) to determine whether it matches.

3. Relative quantification of metabolites

After the above two major steps, the prerequisites for relative quantification are met:

(1) The coordinates of the features ["mzmin", "mzmax", "rtmin", "rtmax"], these coordinates maximize the coverage area of the precursor ion (mz) and reduce the mass-to-charge ratio mz and The impact of fluctuations in retention time RT.

(2) MS2 database matching the features and the identification results of MS2.

(3) For each feature, there are ["mzmin", "mzmax", "rtmin", "rtmax", "metabolite", "adduct"], and one feature can be fully annotated (remarks: "Metabolite" and " "Adduct" information is obtained from reference databases (such as NISTlib, etc.), and "Metabolite" and "adduct" information are set as qualitative processes).

The quantitative method is as follows:

1) Correct the time axis of the sample according to the reference sample information (reference.xml).

2) Integrate the features area of the sample to obtain the relative quantitative results of metabolites.

Under the overall inventive concept of the present invention, the present invention also provides the above-mentioned biological metabolomics data processing method and biological metabolomics data analysis method in vitamins, amino acids, lipids, steroids, aromatic acids, neurotransmitters, pigments , Application in the identification of carbohydrates or short peptides. Since the above-mentioned biological metabolomics data processing method of the present invention is not affected by the test batch, it can continuously accumulate the sample data volume of the characteristic database, thereby also increasing vitamins, amino acids, lipids, steroids, aromatic acids, and neurotransmitters. Quality, pigment, carbohydrate or short peptide identification accuracy and precision.

Further, the present invention also provides a method for detecting vitamins, amino acids, lipids, steroids, aromatic acids, neurotransmitters, pigments, carbohydrates or short peptides. The detection method includes: performing liquid chromatography-mass spectrometry and/or gas chromatography-mass spectrometry on a biological sample to obtain liquid chromatography-mass spectrometry data and/or gas chromatography-mass spectrometry data; using any of the above-mentioned biological metabolomics data processing Method or analysis method of biological metabolomics data process the liquid chromatography-mass spectrometry data and/or gas chromatography-mass spectrometry data of biological samples to obtain data results; and convert vitamins, amino acids, lipids, steroids, and aromas based on the data results The type and content of acids, neurotransmitters, pigments, carbohydrates or short peptides. Similarly, due to the advancement of the biological metabolomics data processing method and analysis method of the present invention, the test results of vitamins, amino acids, lipids, steroids, aromatic acids, neurotransmitters, pigments, carbohydrates or short peptides must also be More precise.

In addition, in order to facilitate the implementation of the above method of the present invention, under the purpose of the present invention, according to a typical embodiment of the present invention, a biological metabolomics data processing device is provided. Biological metabolomics data includes liquid chromatography-mass spectrometry data and/or gas chromatography-mass spectrometry data, liquid chromatography-mass spectrometry data includes primary mass spectrometry data, and gas chromatography-mass spectrometry data includes primary mass spectrometry data; the biological metabolomics data The processing device includes a database generation module that integrates liquid chromatography-mass spectrometry data or gas chromatography-mass spectrometry data of multiple biological samples to form a feature database. The database generation module includes: a time axis correction submodule, a feature peak recognition submodule, and features The database forms a sub-module, where the time-axis correction sub-module is set to arbitrarily select one of the multiple biological samples as a reference sample, and correct the time-axis of other samples one by one according to the time axis of the reference sample; the characteristic peak recognition sub-module setting In order to perform the peak identification processing of the first mass spectrum ion peaks one by one for each sample after calibration, to obtain multiple identification characteristic peaks; and the feature database formation sub-module is set to merge the multiple identification characteristic peaks according to the principle of sample information complementarity Through processing, a feature database of multiple biological samples is obtained.

Using the device of the present invention, first select a sample as the reference sample, and the time axis of other samples are corrected according to this sample, that is, a unified coordinate axis is determined so that the liquid chromatography-mass spectrometry data or gas chromatography-mass spectrometry data of the sample is in time It is comparable; on the corrected time axis, peak identification is performed on the primary mass spectrum ion peak of each sample, and then the peaks are merged using the principle of information complementarity between samples to construct a feature database, which can be realized Super large-scale integration of metabolome data. Since all samples are calibrated on the time axis based on the reference sample, data calibration and data integration can be achieved in batches or individual samples, and are not affected by the test batch. This is suitable for commercial testing.

In an embodiment of the present invention, the mass spectrum data further includes secondary mass spectrum data, and the time axis correction sub-module is further configured to perform retention time correction on the primary mass spectrum data and the secondary mass spectrum data to further improve the accuracy of the mass spectrum data; preferably , Use Obiwarp algorithm for retention time correction, which has the advantages of fast calculation speed and high data processing accuracy.

In an embodiment of the present invention, the feature database forming sub-module includes a data integration unit, and the data integration unit is configured to overlap or adjacent [mzmin, mzmax] regions, and multiple identifications where the [rtmin, rtmax] regions overlap or are adjacent The characteristic peaks are merged into one characteristic peak; preferably, the characteristic database forming sub-module includes a first judgment unit, a second judgment unit, a data integration unit, and a characteristic database forming unit: wherein the first judgment unit is configured to judge multiple identification characteristic peaks If the [mzmin, mzmax] regions overlap, if they overlap, enter the data integration unit; if they do not overlap, further determine whether they are adjacent. If the interval between the [mzmin, mzmax] regions of multiple identifying characteristic peaks is less than the first preset threshold, If it is judged to be adjacent, go to S133; if it is neither overlapping nor adjacent, judge that multiple identification characteristic peaks are independent characteristic peaks; the second judgment unit is set to judge the [rtmin, rtmax ] Whether the regions overlap or are adjacent, if they overlap, enter the data integration unit; if they do not overlap, further determine whether they are adjacent, if the interval between the [rtmin, rtmax] regions of multiple identifying characteristic peaks is less than the second preset threshold, then determine If it is adjacent to each other, enter the data integration unit; if it is neither overlapping nor adjacent, it is determined that multiple identification characteristic peaks are independent characteristic peaks; the data integration unit is set to overlap or adjacent to the [mzmin, mzmax] area, And the [rtmin, rtmax] area overlaps or multiple adjacent identification feature peaks are merged into one feature peak; the database forming module uses the data of all the feature peaks to generate a feature list to obtain the feature database. In this way, the combined characteristic peaks can cover a larger area, even if only one sample is detected, it can be quantified more accurately (effective for metabolites with poor chromatographic peak shapes), and a larger coverage area can be more effective Compatible with subsequent samples, effectively reducing the impact of retention time (RT) drift.

Preferably, the peak recognition algorithm is CentWave algorithm, matchedFilter algorithm or mzMine algorithm, more preferably CentWave algorithm, because this method can improve sensitivity, limit its error, and find the most recognized characteristic peaks most accurately. In this way, peaks are merged on the basis of CentWave algorithm, which can effectively use CentWave to locate the maximum response area. In the present invention, the algorithm parameter setting follows the idea of "improving detection sensitivity as much as possible". Preferably, the parameter setting of the peak recognition algorithm includes: ppm: the resolution of the instrument is adopted; peak width: set to 2-30; noise: Set to 0; SNR: set to 10.

In an embodiment of the present invention, the first preset threshold is set according to instrument parameters, and the second preset threshold is set according to the maximum value of the time deviation in the retention time correction; preferably, the first preset threshold is set It is 0.01~0.015Da, more preferably, the first preset threshold is set to 0.01Da, 0.011Da, 0.012Da, 0.013Da, 0.014Da or 0.015Da, and the second preset threshold is set to 10~15, more preferably Yes, the second preset threshold is set to 10, 11, 12, 13, 14 or 15, to improve the effectiveness of peak merging, thereby improving the accuracy of the feature database.

In an embodiment of the present invention, preferably, the mass spectrometry data further includes secondary mass spectrometry data, and the biological metabolomics data processing device further includes: a peak merging validity verification submodule configured to combine the secondary mass spectrometry data of multiple biological samples Compare to the feature database to assist in judging the effectiveness of peak merging. The higher the ratio of the secondary mass spectrum data to the feature database, the stronger the effectiveness of peak merging. In an embodiment of the present invention, the secondary mass spectrometry data of multiple biological samples are compared to a feature database, wherein when the comparison rate is greater than or equal to the third preset threshold, it is determined that the peak combination is effective; a more preferred embodiment In a more preferred embodiment, the third preset threshold is set to 40%; in a more preferred embodiment, the third preset threshold is set to 50%; in a more preferred embodiment, the third preset threshold is set to 60% In a more preferred embodiment, the third preset threshold is set to 80%.

Under the purpose of the present invention, a biological metabolomics data analysis device is also provided. The analysis device includes a module configured to process biological metabolomics data and a module configured to qualitatively identify metabolites through secondary mass spectrometry data information, wherein the module configured to process biological metabolomics data is the aforementioned biological Metabolomics data processing device. Since the above-mentioned biological metabolomics data processing device of the present invention can not be affected by the test batch, it can continuously accumulate the sample data volume of the characteristic database, thereby continuously improving the qualitative identification of metabolites through the secondary mass spectrum data information accuracy.

According to a typical implementation of the present invention, the module configured to qualitatively identify metabolites through the data information of the secondary mass spectrum includes a standard compound mass-to-charge ratio data acquisition sub-module and a standard compound matching sub-module, wherein the standard compound mass-to-charge ratio The data acquisition sub-module is set to acquire the mass-to-charge ratio data of each standard compound; the standard compound matching sub-module is set to randomly select a characteristic value in the characteristic database obtained after the biological metabolomics data processing, and find the corresponding characteristic value All the mass-to-charge ratio data of the secondary mass spectrum, according to all the mass-to-charge ratio data of the secondary mass spectrum, find the matching standard compound; the integration qualitative sub-module is set to correspond to a characteristic value selected in the standard compound matching sub-module All the mass-to-charge ratio data of the second-level mass spectra of, take the mass-to-charge ratio data of the matched standard compound found in the standard compound matching sub-module as the other side, score the similarity between the two, and calculate the point integral. The metabolites are qualitatively based on the integral value. This method draws on the parameter setting of the knn algorithm and the merging method of the density algorithm, which can effectively avoid the problem that the median is not representative, and the operation is simple.

Preferably, the integral qualitative sub-module is set to calculate the median of the similarity between each standard compound in the multiple standard compounds on the match and the data of multiple secondary mass spectrometry, and select the compound with the largest median; more preferably, according to the compound Whether the median of is greater than the cut-off value, judge whether it matches. The above algorithm not only includes the "representative" MS2, but also adds various possible MS2 of the compound, which increases the matching degree with the standard compound.

In the present invention, the mass-to-charge ratio data of the standard compound is obtained from an existing database, for example, the database includes NISTlib, HMDB or METLIN.

In a typical embodiment of the present invention, the analysis device further includes a module configured to quantify biological metabolites. After the above-mentioned data processing and characterization, a unified time axis was determined, the retention time was corrected, and a feature database with rich data volume was obtained, so that the coverage area of the precursor ion (mz) was increased as much as possible. Reduce the impact of fluctuations in the mass-to-charge ratio mz and retention time RT, and improve the accuracy of biological metabolite quantification. Preferably, the module configured to quantify biological metabolites includes a time axis correction sub-module and a biological metabolite relative quantification sub-module, wherein the time axis correction sub-module is set to correct the time axis of the sample to be quantified according to the time axis of the reference sample; The relative quantification of biological metabolites sub-module is set to integrate the corresponding characteristic regions of the samples to be quantified in the established characteristic database to obtain the relative quantitative results of biological metabolites.

In the following, the beneficial effects of the present invention will be further described in conjunction with embodiments.

Taking 101 dried blood slice samples as an example, using the technical solution of the present invention (Example 1) and the method of the prior art (Comparative Example 1) to integrate the metabolome data of these 101 samples, and perform qualitative and quantitative analysis at the same time ,details as follows.

Example 1

1. Build the features database

1) Determine a unified coordinate axis to make the samples comparable in time. A fixed sample is selected from all samples as a reference, and the time axis of other samples is corrected according to this sample, which is reference.xml.

2) The new sample is first corrected for retention time (RT). In this step, the Obiwarp algorithm is used to correct the retention time (RT) of the primary mass spectrum data (MS1) and the secondary mass spectrum data (MS2). The sample retention time correction in this embodiment is shown in Figure 3 (Note: the horizontal axis is the retention time RT (unit: s), and the vertical axis is the time the retention time of the sample deviates from the reference sample (unit: s), also called retention Time deviation. The horizontal line is the reference sample, and the curve is the other samples.

3) On the corrected time axis, use the CentWave algorithm to identify peaks (findPeaks) for each sample, and find the most peaks most accurately. The noise caused by high sensitivity and the problem of the same peak being divided into two peaks caused by the strict ppm setting are handled by the complementary sample information.

The algorithm parameter setting follows the principle of "improving the detection sensitivity as much as possible". The specific settings in this embodiment are as follows:

ppm: 10

Peakwidth: 2～30

Noise: 0

Snthresh: 10

Results: Each sample has about 3,600 to 5,000 peaks, and 101 samples have a total of 431,695 peaks.

4) According to the principle of sample information complementarity, the peaks of 101 samples are combined to generate a unified coordinate, that is, a feature database. The specific processing is as follows (see Figure 2):

For peak1, peak2,...,peakn from 101 samples, make the following judgments:

①Whether the [mzmin,mzmax] regions in the 101 samples overlap or are adjacent, if they overlap, go to step ③; if they do not overlap, further judge whether they are adjacent, if peak m+1, peak m+2,..., peak m The interval of the [mzmin, mzmax] area of +a is less than the first preset threshold 0.015Da, it is judged to be adjacent, and enter ③; if it is neither overlapping nor adjacent, it is judged that multiple identification feature peaks are independent features peak;

②Whether the [rtmin,rtmax] regions of the 101 samples overlap or are adjacent, if they overlap, go to step ③; if they do not overlap, judge whether they are adjacent: set the second preset threshold absRt=15, if peak m+1, peak m+2,..., peak m+a If the interval of the RT region is less than absRt, it is judged to be adjacent, and proceed to step ③; if it is neither overlapping nor adjacent, it is judged that multiple identified characteristic peaks are independent features peak;

③If the relationship of peak m+1, peak m+2,..., peak m+a satisfies the above two conditions of overlap/adjacent and overlap/adjacent at the same time, then it is determined that peak m+1, peak m+2,..., peak m+a belongs to the same peak. The coordinates of the new peak are taken as the union of a to generate a feature list. Among them, n and a are independently valued in positive integers, m is valued in 0 and positive integers, and m<n.

Result: After merging, 23799 features (feature database) are generated.

4) Compare MS2 to the merged features (rectangular area determined by ["mzmin", "mzmax", "rtmin", "rtmax"]). There are 346,518 MS2s in 101 samples, of which 279,568 are compared to features, with a comparison rate of 80.68%, while the comparison rate of MS2 for a single sample is only about 50%. A total of 6916 features have corresponding MS2, and the number of MS2 owned by these features ranges from 1 to 2272.

5) According to the known information in the feature database, the metabolome data of 101 samples are sequentially integrated.

The results of the integration are shown in Table 1:

Table 1

2. Metabolite identification

Take one of the features (No. FT08341) as an example, its mz is [352.158899,352.168942] and RT is [167.3529,189.8049].

The metabolite identification steps are as follows:

1) Obtain the mass-to-charge ratio mz of the standard compound. Each standard compound contains 18 ionization forms, as shown in Table 2 (Types of ionization forms of standard compounds). When testing the sample, one or more ionized forms of the compound will be obtained. Each compound contains one or more ionized forms. Table 3 lists 5 standard compounds (S0001-S0005) and their corresponding 5 ionized forms. mz.

Table 2

1	M+	7	(M-H+2Na)+	13	(M+CH3CN+Na)+
2	(M+H)+	8	(M-2H+3Na)+	14	(2M+H)+
3	(M+H-H2O)+	9	(M+K)+	15	(2M+NH4)+
4	(M+H-2H2O)+	10	(M-H+2K)+	16	(2M+Na)+
5	(M+NH4)+	11	(M-2H+3K)+	17	(2M+K)+
6	(M+Na)+	12	(M+CH3CN+H)+	18	(M+CH3COO+2H)

table 3

To	M+	(M+H)+	(M+H-H2O)+	(M+H-2H2O)+	(M+NH4)+
S0001	74	75.01	57	38.99	92.03
S0002	112	113	95.02	77.01	130.1
S0003	116	117	99.01	81	134
S0004	116	117	99.01	81	134
S0005	117.1	118.1	100.1	82.04	135.1

Note: S0001 is the number of the compound, M+ is the ionized form, 74 and so on are mz.

2) Find the same standard compound as the parent ion mz of MS2

Specific process:

a. Find all MS2s matching FT08341, a total of 35;

b. Calculate the median mz of the 35 MS2 precursor ions, name it mzmed, and calculate mzmed=352.1652;

c. Search for standard compounds that are the same as the mzmed of these 35 MS2s. A total of 16 are found. The ionized form, mass-to-charge ratio mz and intensity of these standard compounds are shown in Figure 4 (close to the mz of feature (FT08341) MS2) The MS2 of the standard compound, note n01701, etc. are the compound number, (M+) etc. are the ionized form of the compound).

3) Comparing the similarities between the 35 MS2 obtained in the experiment and the 16 standard compounds on the match, see Figure 5. Among them, the similarity between MS2 of FT08341 and n01696 is relatively high (most of the similarities between MS2 and n01696 are greater than 0.8, and the average similarity is the largest).

Note: In Figure 5, the horizontal axis is the standard compound, and the vertical axis is 35 MS2 of FT08341. For example, the first column is the similarity between 35 MS2 and n01701MS2.

4) Synthesize the results of 35 MS2s and select the most matching standard compound for metabolite identification.

Calculate the median of similarity between these 16 standard compounds and 35 MS2, as shown in Table 4. The compound n01696 with the largest median was selected. Since the median of 0.890 of the compound was greater than the specified value (cutoff=0.5), n01696 was identified as a FT08341 matched compound. The result is also consistent with the result in Figure 5.

Table 4

Standard compound	n01701	n01694	n01696	n01835	n01577	n01578	n01579	n01440
median	0.041	0.211	0.890	0.000	0.000	0.005	0.003	0.026
Standard compound	n01444	n01320	L0194	n00528	n01419	n01420	n01421	n01423
median	0.150	0.000	0.000	0.012	0.006	0.494	0.008	0.004

The designated value can be determined as 0.5 to 1 according to actual conditions, and the designated value in this embodiment is cutoff=0.5.

3. Relative quantification of metabolites in a single sample

1) Correct the time axis of the sample according to reference.xml.

2) Integrate the features area of the sample to determine the relative quantitative value of metabolites.

table 5

Table 5 shows the relative quantitative results (partial) of metabolites. GXP104, GX107, etc. are sample names, score is the matching score of the compound, and metabolite is the identified compound.

For example, the name of the metabolite matched by FT08341 is Phe-Trp (the name of the metabolite is obtained in the database of standard compounds, the database is mainly NISTlib, and the database can also be replaced with HMDB, METLIN and other public databases), and the score is 0.89. , The credibility is high. In the GXP104 sample, the relative quantitative value of the metabolite is 8495.393221, and the relative quantitative value in GXP107 is 5096.885985.

As another example, as shown in Table 5, the relative quantitative value of compound FT02707 in sample GXP104 is 12386.06788; the relative quantitative value of compound FT05421 in sample GXP104 is 2252.548371.

It can be seen that using the technical scheme of the present invention can accurately perform relative quantification of metabolites.

In the above embodiments, liquid chromatography-mass spectrometry data is processed. Those skilled in the art can understand that gas chromatography-mass spectrometry data can also be processed by this method, and the same technical effect can be obtained.

Comparative example 1

1) Convert the original files of 101 dried blood tablets into mzXML format;

2) Use the obiwarp method of XCMS to correct the retention time and correct the time of each scan. The relevant parameters are set as:

ppm: 25

Peakwidth: 4～10

Noise: 10

Snthresh: 3

3) Use the CentWave method of XCMS to identify ion peaks;

4) The peaks of a single mzXML are obtained in the above 3 steps, and the peaks exist in the form of ["mzmin","mzmax","rtmin","rtmax","into","maxo"], where ["mzmin","mzmax ","Rtmin","rtmax"] are the coordinates of peaks, ["into","maxo"] are quantitative information (integral and maximum);

5) A series of features are obtained by aligning and grouping the peaks, ensuring that the ["mzmin", "mzmax", "rtmin", "rtmax"] area does not overlap. Here mzmin is the minimum value of the median mz of multiple samples, and mzmax is the maximum value of the median mz of multiple samples. A total of 4289 features are obtained from 101 samples;

6) Missing value filling: Integrate the relevant area of mzXML according to the unified coordinates. Missing value filling is to integrate the area according to the found coordinates.

The treatment results of Example 1 and Comparative Example 1 are compared as follows:

1. Example 1 found 23799 features, 21273 feature non-missing values, comparative example 1 analyzed 4289 features, feature non-missing values 4042. It shows that the technical solution of the present invention can find more features, and the number of feature missing values is less. FIG. 6 shows the distribution of the number of feature missing values in Example 1. There are 101 samples in total, and 85% of the feature missing values are less than 20.

2. Figure 7 shows a comparison of the coefficient of variation (CV, standard deviation divided by the mean) between the samples of Example 1 and Comparative Example 1. Among them, in Figure 7, the median line (the first straight line from left to right in the coordinate system (parallel to the ordinate)) represents the median, and the quartile line (the second from left to right in the coordinate system) The straight line (parallel to the ordinate) represents the upper quartile (75%), that is, there is 50% on the left side of the median line (the first straight line from left to right in the coordinate system (parallel to the ordinate)) The CV value of the features is less than the value corresponding to the median line. On the left side of the quartile line (the second straight line from left to right in the coordinate system (parallel to the ordinate)), there are 75% of the features’ CV value Less than the value corresponding to the quartile line (the second straight line from left to right in the coordinate system (parallel to the ordinate)). It can be seen from Figure 7 that the coefficient of variation (CV) between samples in Example 1 is smaller, because the feature database is constructed in Example 1, which makes the feature in Example 1 more stable and excludes the peak time (RT) The difference maximizes the abundance of the molecular ion (mz) itself. The consistency of QC samples has also improved.

FIG. 8 shows the PCA results of Example 1 and Comparative Example 1 showing that the consistency between the samples of Example 1 is better than that of Comparative Example 1. On the one hand, the proportion that can be explained by PC1 and PC2 in Example 1 is greatly increased. On the other hand, the distinction between experimental samples (solid circles) and QC samples (dashed circles) is also more obvious.

In addition, the beneficial effect obtained in Example 1 is also manifested in the full use of the MS2 information of the sample. On the one hand, it greatly improves the identification rate of MS2. On the other hand, it brings about a benefit that an MS2 database is generated, which can be used to evaluate new MS2 similarity algorithm. These effects are mainly due to:

1) The same feature has MS2 from multiple samples, which covers multiple fragmentation methods of the precursor ion and improves the matching efficiency with standard compounds, so more metabolites can be identified.

2) Compare multiple MS2s with standard compounds. This method effectively avoids the problem of a single MS2 matching multiple compounds and reduces false positives. Figure 9 shows a comparison of the number of metabolites identified in Example 1 and Comparative Example 1.

3) The correlation between multiple MS2 belonging to the same feature can assist in judging the effectiveness of the peaks merging algorithm.

Furthermore, FIG. 10 shows the mz and RT distributions of MS2 precursor ions corresponding to FT08341 in Example 1. ① If mz and RT are distributed in a very narrow range, it can be concluded that they belong to the same parent ion, so they can be used to evaluate the MS2 similarity algorithm. ② If the range of mz and RT (mainly RT) is relatively wide, given the MS2 similarity algorithm, the corresponding MS2 can be used to evaluate whether the precursor ions are the same, which can assist in judging the rationality of peaks merging.

Figure 11 shows the similarity of 35 MS2 spectra of FT08341 in Example 1. The similarity comparison of multiple MS2s of the same feature can be used to help judge the effect of combining peaks into features.

From the above description, it can be seen that the above-mentioned embodiments of the present invention achieve the following technical effects:

1. In terms of time axis correction, a fixed reference sample can ensure that subsequent samples are comparable in time.

2. This method merges peaks based on the CentWave algorithm, and effectively utilizes CentWave's positioning of the maximum response area.

3. The parameter setting of the knn algorithm and the merging method of the density algorithm are used for reference, and the problem that the median is not representative is effectively avoided, and the operation is simple.

4. The combined peak can cover a larger area, even if only one sample is tested, it can be more accurately quantified (effective for metabolites with poor chromatographic peak shape), and a larger coverage area can be more effectively compatible Subsequent samples effectively reduce the impact of retention time (RT) drift.

5. After the database is established, the efficiency of sample analysis is improved, and subsequent samples are comparable in time, so there is no need to roll back the samples, which improves commercial availability.

The foregoing descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. For those skilled in the art, the present invention can have various modifications and changes. Any modification, equivalent replacement, improvement, etc., made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Industrial applicability

Through the technical methods of this application, the present invention has at least the following beneficial effects:

Applying the technical solution of the present invention, by constructing a feature database, unifying the time axis, using the principle of information complementarity between samples for peak merging, etc., ultra-large-scale metabolome data integration can be realized, and batches can be realized Or the data correction and data integration of a single sample is not affected by the test batch, and it is suitable for commercial testing.

The invention constructs a feature database, fixes a reference sample, and unifies the time axis, which can ensure that subsequent samples are comparable in time, so that the metabolome data processing process can effectively use sample information complementation between different batches, and effectively improve metabolism. Object detection repeatability and coverage.

The present invention performs peak merging processing in the process of constructing a feature database, and the merged peaks can cover a larger area, so that quantification can be performed more accurately when only one sample is detected, even if the chromatographic peak shape is not good Metabolites still have a good effect, and produce a larger coverage area, which makes it more compatible with subsequent samples, and effectively reduces the impact of retention time (RT) drift.

After the feature database is established, the present invention effectively improves the analysis efficiency of samples, so that subsequent samples are comparable in time, and the samples do not need to be rolled back, and can be widely used in business.

Claims (31)

1. A method for processing biological metabolomics data, wherein the biological metabolomics data includes liquid chromatography-mass spectrometry data or gas chromatography-mass spectrometry data, the liquid chromatography-mass spectrometry data includes primary mass spectrometry data, and the gas phase Chromatography-mass spectrometry data includes primary mass spectrometry data; the biological metabolomics data processing method includes the step of integrating liquid chromatography-mass spectrometry data or gas chromatography-mass spectrometry data of a plurality of biological samples to form a feature database. The steps include:

   S11, arbitrarily selecting one of the plurality of biological samples as a reference sample, and correcting the time axes of other samples one by one according to the time axis of the reference sample;

   S12, for each sample after calibration, perform peak identification processing of the ion peaks of the primary mass spectrum one by one to obtain multiple identification characteristic peaks; and

   S13: According to the principle of complementary sample information, the multiple identification characteristic peaks are combined to obtain a characteristic database of the multiple biological samples.
2. The biological metabolomics data processing method according to claim 1, in said S13: if the [mzmin, mzmax] regions of the multiple identification characteristic peaks overlap or are adjacent, and the [rtmin, rtmax] regions overlap or are in phase Adjacent, merge the multiple identification characteristic peaks into one characteristic peak.
3. The biological metabolomics data processing method according to claim 2, wherein S13 comprises:

   S131: Determine whether the [mzmin, mzmax] regions of the multiple identification feature peaks overlap or are adjacent, if they overlap, go to S133; if they do not overlap, further determine whether they are adjacent, if the [mzmin , Mzmax] area is less than the first preset threshold, it is determined to be adjacent, and enter S133; if it is neither overlapping nor adjacent, it is determined that the multiple identification characteristic peaks are independent characteristic peaks;

   S132: Determine whether the [rtmin, rtmax] regions of the multiple identification feature peaks overlap or are adjacent, if they overlap, go to S133; if they do not overlap, further determine whether they are adjacent, if the [rtmin, rtmax] regions of the multiple identification feature peaks , Rtmax] The interval of the region is less than the second preset threshold, it is determined as being adjacent, and entering S133; if it is neither overlapping nor adjacent, then determining that the plurality of identification characteristic peaks are independent characteristic peaks;

   S133: If the multiple identification characteristic peaks simultaneously satisfy the overlapping or adjacent conditions in S131 and the overlapping or adjacent conditions in S132, merge the multiple identification characteristic peaks into one characteristic peak;

   S134: Use the data of all the characteristic peaks to generate a characteristic list to obtain the characteristic database.
4. The biological metabolomics data processing method according to claim 3, wherein the first preset threshold is set according to instrument parameters, and the second preset threshold is set according to the maximum value of the time deviation in the retention time correction ；

   Preferably, the first preset threshold is set to 0.01 to 0.015 Da, and the second preset threshold is set to 10 to 15.
5. The biological metabolomics data processing method according to claim 3, wherein the mass spectrometry data further comprises secondary mass spectrometry data, and the S13 further comprises:

   S135. Compare the secondary mass spectrometry data of the multiple biological samples to the feature database generated in S134, wherein when the comparison rate is greater than or equal to a third preset threshold, it is determined that the peak combination is effective;

   Preferably, the third preset threshold is set to 40%; more preferably, the third preset threshold is set to 50%; more preferably, the third preset threshold is set to 60%; More preferably, the third preset threshold is set to 80%.
6. The biological metabolomics data processing method according to claim 1, wherein the mass spectrometry data further comprises secondary mass spectrometry data, and the S11 further comprises performing retention time correction on the primary mass spectrometry data and the secondary mass spectrum data;

   Preferably, the Obiwarp algorithm is used for retention time correction.
7. The biological metabolomics data processing method according to claim 1, wherein the algorithm for peak identification is CentWave algorithm, matchedFilter algorithm or mzMine algorithm.
8. The biological metabolomics data processing method according to claim 7, wherein the parameter settings of the peak recognition algorithm include: ppm: the resolution of the instrument; peak width: set to 2-30; noise: set to 0; Noise ratio: set to 10.
9. The biological metabolomics data processing method according to claim 1, wherein the biological sample comprises human or animal body fluids, tissues or cells, plant roots, stems, leaves, fruits or seeds, or microbial cell culture fluid; wherein The body fluid includes urine, blood, saliva, cerebrospinal fluid or amniotic fluid, the tissue includes organ tissue, muscle tissue or tumor tissue, and the cell includes stem cells, somatic cells, tumor cells or microbial cells.
10. A method for analyzing biological metabolomics data, including the steps of biological metabolomics data processing and qualitative identification of metabolites through secondary mass spectrometry data information, wherein the biological metabolomics data processing adopts the steps of claim 1. The biological metabolomics data processing method described in any one of to 9 is performed.
11. According to the analysis method of claim 10, the step of qualitatively identifying metabolites through the data information of secondary mass spectrometry comprises:

    S21, obtain the mass-to-charge ratio data of each standard compound;

    S22. Select a characteristic value arbitrarily in the characteristic database obtained after the biological metabolomics data processing, and find all the mass-to-charge ratio data of the secondary mass spectrum corresponding to the characteristic value, and according to the mass-to-charge ratio of all the secondary mass spectra Data, find a standard compound that matches it;

    S23, taking the mass-to-charge ratio data of all the MS mass spectra corresponding to the one characteristic value selected in S22 as one side, and taking the mass-to-charge mass spectra of the matched standard compound found in S22 The comparison data is the other side, score the similarity of the two, calculate the points, and qualitative metabolites based on the integral value.
12. The analysis method according to claim 11, said S23 comprising: calculating the median of the similarity between each standard compound in the plurality of standard compounds on the matching and the data of a plurality of secondary mass spectrometry, and selecting the one with the largest median Compound

    Preferably, according to whether the median of the compound is greater than the cut-off value, it is judged whether it matches.
13. The analysis method according to claim 11, wherein the mass-to-charge ratio data of the standard compound is obtained from an existing database, and the database includes NISTlib, HMDB or METLIN.
14. The analysis method according to claim 10, further comprising a step of quantifying biological metabolites.
15. The method according to claim 14, wherein the step of quantifying biological metabolites comprises:

    S31: Correct the time axis of the sample to be quantified according to the time axis of the reference sample;

    S32: Integrating the corresponding characteristic regions of the sample to be quantified in the established characteristic database to obtain a relatively quantitative result of biological metabolites.
16. The method for processing biological metabolomics data according to any one of claims 1 to 9, and the method for analyzing biological metabolomics data according to any one of claims 10 to 15 include vitamins, amino acids, lipids, steroids, Application in the identification of aromatic acids, neurotransmitters, pigments, carbohydrates or short peptides.
17. A method for detecting vitamins, amino acids, lipids, steroids, aromatic acids, neurotransmitters, pigments, carbohydrates or short peptides, which is characterized in that it comprises: performing liquid chromatography-mass spectrometry and/or gas chromatography- on biological samples Mass spectrometry detection to obtain liquid chromatography-mass spectrometry data and/or gas chromatography-mass spectrometry data; using the biological metabolomics data processing method according to any one of claims 1 to 9 or any one of claims 10 to 15 The method for analyzing biological metabolomics data processes liquid chromatography-mass spectrometry data and/or gas chromatography-mass spectrometry data of the biological sample to obtain data results; and converts the vitamins and amino acids according to the data results , Lipids, steroids, aromatic acids, neurotransmitters, pigments, carbohydrates or short peptides.
18. A biological metabolomics data processing device, wherein the biological metabolomics data includes liquid chromatography-mass spectrometry data or gas chromatography-mass spectrometry data, the liquid chromatography-mass spectrometry data includes primary mass spectrometry data, and the gas phase Chromatography-mass spectrometry data includes primary mass spectrometry data; the biological metabolomics data processing device includes a database generation module that integrates liquid chromatography-mass spectrometry data or gas chromatography-mass spectrometry data of multiple biological samples to form a feature database, The database generation module includes:

    The time axis correction sub-module is configured to arbitrarily select one of the plurality of biological samples as a reference sample, and correct the time axes of other samples one by one according to the time axis of the reference sample;

    The characteristic peak recognition sub-module is configured to perform peak recognition processing of the ion peaks of the primary mass spectrometer one by one for each sample after calibration to obtain multiple characteristic peaks; and

    The feature database forming sub-module is configured to merge the multiple identification feature peaks according to the principle of sample information complementarity to obtain feature databases of the multiple biological samples.
19. The biological metabolomics data processing device according to claim 18, the feature database forming sub-module includes a data integration unit, and the data integration unit is configured to overlap or adjacent to [mzmin, mzmax] regions, and [rtmin, rtmax] The overlapping or adjacent multiple identification characteristic peaks are merged into one characteristic peak.
20. The biological metabolomics data processing device according to claim 19, the characteristic database forming sub-module includes a first judgment unit, a second judgment unit, the data integration unit and a characteristic database forming unit:

    Wherein, the first determining unit is configured to determine whether the [mzmin, mzmax] regions of the multiple identification characteristic peaks overlap, if they overlap, enter the data integration unit; if they do not overlap, further determine whether they are adjacent, if If the interval between the [mzmin, mzmax] regions of the multiple identification characteristic peaks is less than the first preset threshold, it is determined to be adjacent, and the process proceeds to S133; if neither overlap nor adjacent, then the multiple identification characteristic peaks are determined Are independent characteristic peaks;

    The second determining unit is configured to determine whether the [rtmin, rtmax] regions of the multiple identification characteristic peaks overlap or are adjacent, if they overlap, enter the data integration unit; if they do not overlap, further determine whether they are adjacent, If the interval between the [rtmin, rtmax] regions of the plurality of identification characteristic peaks is less than the second preset threshold, it is determined to be adjacent and enters the data integration unit; if it is neither overlapping nor adjacent, it is determined that the Multiple identification characteristic peaks are independent characteristic peaks;

    The data integration unit is configured to merge multiple identification characteristic peaks that overlap or be adjacent to [mzmin, mzmax] regions, and to merge or adjacent [rtmin, rtmax] regions into one characteristic peak;

    The feature database forming unit is configured to generate a feature list using data of all feature peaks to obtain the feature database.
21. The biological metabolomics data processing device according to claim 20, wherein the first preset threshold is set according to instrument parameters, and the second preset threshold is set according to the maximum value of the time deviation in the retention time correction ；

    Preferably, the first preset threshold is set to 0.01 to 0.015 Da, and the second preset threshold is set to 10 to 15.
22. The biological metabolomics data processing device according to claim 20, wherein the mass spectrometry data further comprises secondary mass spectrometry data, and the biological metabolomics data processing device further comprises:

    The peak merging validity verification sub-module is configured to compare the secondary mass spectrum data of the multiple biological samples to the feature database, wherein when the comparison rate is greater than or equal to a third preset threshold, it is determined that the peak merging is valid ；

    Preferably, the third preset threshold is set to 40%; more preferably, the third preset threshold is set to 50%; more preferably, the third preset threshold is set to 60%; More preferably, the third preset threshold is set to 80%.
23. The biological metabolomics data processing device according to claim 18, wherein the mass spectrometry data further comprises secondary mass spectrometry data, and the time axis correction submodule is further configured to compare the primary mass spectrum data and the secondary mass spectrum data Perform retention time correction;

    Preferably, the Obiwarp algorithm is used for retention time correction.
24. The biological metabolomics data processing device according to claim 18, wherein the algorithm for peak recognition is CentWave algorithm, matchedFilter algorithm or mzMine algorithm.
25. The biological metabolomics data processing device according to claim 24, the parameter settings of the peak recognition algorithm include: ppm: the resolution of the instrument; peak width: set to 2-30; noise: set to 0; Noise ratio: set to 10.
26. A biological metabolomics data analysis device, comprising a module configured to process biological metabolomics data and a module configured to qualitatively identify metabolites through secondary mass spectrometry data information, wherein the setting is biological metabolomics The data processing module is the biological metabolomics data processing device according to any one of claims 18 to 25.
27. The analysis device according to claim 26, wherein the module configured to qualitatively identify metabolites through secondary mass spectrometry data information comprises:

    The standard compound mass-to-charge ratio data acquisition sub-module is set to acquire the mass-to-charge ratio data of each standard compound;

    The standard compound matching submodule is set to randomly select a characteristic value in the characteristic database obtained after biological metabolomics data processing, and find all the mass-to-charge ratio data of the secondary mass spectrometry corresponding to the characteristic value, according to the All the mass-to-charge ratio data of the secondary mass spectrometer, find the matching standard compound;

    The integral qualitative sub-module is set to take all the mass-to-charge ratio data of the secondary mass spectrometer corresponding to the one characteristic value selected in the standard compound matching sub-module as one side, and find it in the standard compound matching sub-module The mass-nucleus ratio data of the secondary mass spectrum of the matched standard compound is the other party, and the two are scored for similarity, the points are calculated, and the metabolites are qualitatively based on the integral value.
28. The analysis device according to claim 27, wherein the integration and qualitative sub-module is configured to calculate the median of the similarity between each standard compound in the plurality of standard compounds on the match and the data of a plurality of secondary mass spectrometry, and select the median The largest number of compounds;

    Preferably, according to whether the median of the compound is greater than the cut-off value, it is judged whether it matches.
29. The analysis device according to claim 27, wherein the mass-to-charge ratio data of the standard compound is obtained from an existing database, and the database includes NISTlib, HMDB or METLIN.
30. The analysis device according to claim 26, further comprising a module configured to quantify biological metabolites.
31. The device according to claim 30, wherein the module configured to quantify biological metabolites comprises:

    The time axis correction sub-module is set to correct the time axis of the sample to be quantified according to the time axis of the reference sample;

    The relative quantification of biological metabolites sub-module is configured to integrate the corresponding characteristic regions of the sample to be quantified in the established characteristic database to obtain the relative quantitative results of biological metabolites.

Images