An equi-rectangular region of interest (eROI) method for gas chromatography-high resolution mass spectrometry fingerprint data compression
By using the eROI method to perform isochronization on GC-HRMS data, the problems of large data volume and complexity are solved, enabling rapid and accurate qualitative and quantitative analysis, which is suitable for GC-HRMS analysis of complex samples.
Patent Information
- Application Number
- CN202510033913.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-08
- Publication Date
- 2026-07-10
AI Technical Summary
Existing GC-HRMS fingerprint data processing methods suffer from problems such as large data volume and complexity, making it difficult to accurately obtain qualitative and quantitative information. Furthermore, existing algorithms are not sensitive to embedded peaks, require preprocessing such as background subtraction, and require m/z column alignment of multiple sets of data.
The isotropic region of interest (eROI) method is adopted. The original data is processed by normalized sample testing, EIC extraction, column isotropicization and row isotropicization, and converted into an equal row and column matrix. Then, the cocurrent peak deconvolution is performed in combination with the classical multivariate resolution method.
It achieves data compression, improves processing speed, ensures the accuracy of m/z readings, maintains the same retention time, and can quickly obtain accurate qualitative and quantitative data, making it suitable for GC-HRMS analysis of complex samples.
Smart Images

Figure CN122361687A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of non-targeted gas chromatography-high resolution mass spectrometry (GC-HRMS) analysis. Specifically, it relates to an isometric region of interest (eROI) method for GC-HRMS fingerprint data compression, which combines it with classical multivariate resolution methods to accurately obtain qualitative and quantitative information of complex fingerprints. Background Technology
[0002] Compared to in-situ mass spectrometry and mass spectrometry imaging, chromatography-mass spectrometry (GC-MS) is more widely used in metabolomics research due to its high efficiency in separation and sensitive detection. Non-targeted metabolomics, in particular, requires no prior knowledge and can detect and semi-quantitatively quantify the broadest range of metabolites in a sample in an unbiased manner, broadening the horizons for discovering new biomarkers or metabolic pathways. Triple quadrupole mass spectrometry (qMS) or time-of-flight mass spectrometry (ToF MS) coupled with one-dimensional or even multi-dimensional gas chromatography has long been widely used for metabolomics analysis of plant, food essential oils, or biological samples. In this process, metabolite annotation and quantitative analysis have laid a solid foundation for subsequent discovery of plant metabolic pathways, pharmacodynamic bases, and quality markers. However, when faced with samples containing dozens or even hundreds of components with vastly different concentrations, these instruments generally face challenges such as "low mass resolution, narrow linear range, and numerous mismatched peaks." Encouragingly, gas chromatography-orbitrap high resolution mass spectrometry (GC-Orbitrap HRMS) offers higher sensitivity, resolution, and accuracy, along with a wider linear detection range, making it a promising technology for analyzing plant, food, or other metabolomics samples.
[0003] Non-targeted scanning leads to massive storage and complex structures in GC-HRMS. Therefore, chemometric approaches have become essential for omics fingerprint analysis and its interpretation. However, achieving more accurate metabolite detection, qualitative and quantitative analysis through "co-current peak cluster" deconvolution remains challenging. Unlike electrospray ionization (ESI), electron ionization (EI) is a hard ionization method where energy transfer results in the generation of numerous fragment ions. This means that extracting pure mass spectra of each target component from GC-MS data can significantly aid in understanding the fragmentation patterns and structural annotation of target compounds. While numerous multivariate resolution and multidimensional correction algorithms exist for deconvolution in GC-qMS data, they are not suitable for GC-HRMS data. Currently, only methods or platforms such as ROI-MCR, ADAP-GC 4.0, and MS-DIAL can perform deconvolution of GC-HRMS fingerprints. These methods have certain limitations, such as requiring background subtraction and other data preprocessing before deconvolution can be performed based on pure component region information, and being insensitive to embedded peaks. Specifically, using the conventional ROI method for GC-HRMS fingerprint compression results in a data structure where the m / z number does not match the actual number of columns, and requires m / z column alignment across multiple datasets. Therefore, there is an urgent need for more data compression methods that can be combined with classical multivariate resolution methods to accurately obtain qualitative and quantitative information from GC-HRMS fingerprints. These iterative or non-iterative multivariate resolution methods, including HELP, AMWFA, and MCR-ALS algorithms, each have their advantages and disadvantages, and the appropriate method should be selected based on different scenario types. Summary of the Invention
[0004] The purpose of this invention is to overcome the drawbacks of existing GC-HRMS fingerprints, such as massive quantization and complexity, by providing a data compression method. This invention, through its method, can obtain an equal-row, equal-column matrix with performance comparable to the original data, thus preparing for confluence peak deconvolution.
[0005] (1) Sample Testing: ① Obtain the characteristic component extract of the test sample using standardized procedures. Its composition and structure should truly represent the components within the sample. ② If the test sample is a biological sample, a quantitative sample should be weighed, derivatized, and then subjected to GC-Orbitrap HRMS testing. Methodological investigation should be conducted to ensure that the GC peak separation is as complete as possible, with appropriate intensity and reasonable distribution. ③ If the component extract is a plant or food essential oil or other easily (semi) volatile liquid, the test sample should be diluted with an equal volume of n-hexane, injected appropriately, and the chromatographic column optimized to ensure that the GC peak separation is as complete as possible, with appropriate intensity and reasonable distribution. The above test data should have good reproducibility and strong characteristic properties.
[0006] (2) EICs Extraction: ① Convert the raw data into a common format file using software. ② eROI reads the common file (e.g., mzML) and iterates through each mass spectrum to extract retention time, mass-to-charge ratio (m / z), and corresponding peak intensity. ③ Threshold the peak intensities, setting intensities below the threshold to zero, and reduce the m / z value at each retention time to a low-resolution value, grouping them and summing the intensities.
[0007] (2) Column Equalization: ① eROI stores the results in a dictionary, where the key is the retention time and the value is a dictionary containing the sum of peak intensities corresponding to each m / z range. ② eROI sorts the column labels according to the m / z range, so that each m / z falls into a specific position, and writes the results to an Excel file.
[0008] (3) Row alignment: ① eROI reads the Excel file above and extracts the data in the first column. ② The data is grouped according to the integer part of the retention time, and interpolation is performed in each integer part. The entire data is traversed to complete the row alignment interpolation, and the inserted data is saved to a new Excel file.
[0009] (4) eROI-MCR: Convert the Excel file processed above into a mat file (equal row and column matrix), and combine it with the classic "multivariate resolution (MCR)" or "multidimensional correction" algorithm for deconvolution of "common flow peaks" and obtain accurate qualitative or quantitative data.
[0010] The beneficial effects of this invention are:
[0011] Compared to existing ROI technologies, the eROI solution provided by this invention has the following advantages:
[0012] First, both involve compressing the raw data, but eROI's EICs extraction and m / z grouping will result in a loss of accuracy. It is precisely this loss of accuracy that allows the peak intensity corresponding to each m / z value to be placed in the appropriate dictionary position within the column equidistant arrangement, ensuring accurate m / z readings in the mass spectrometer.
[0013] Second, eROI processing speed is much faster.
[0014] Third, after row equalization, the retention time remained unchanged compared to the original data. Attached Figure Description
[0015] Figure 1 AeROI involves three steps to generate a matrix with uniform rows and columns; BeROI-MCR is used for deconvolution of co-eluting peaks.
[0016] Figure 2A biological sample is derivatized and then subjected to GC-Orbitrap HRMS testing; B eROI converts the raw data into a matrix format.
[0017] Figure 3 A. Non-targeted volatile metabolomics fingerprint of Cyperus rotundus obtained by GC-Orbitrap HRMS; BeROI applied to the peak cluster at 33.75–33.90 min; C. Selective ion detection map of the embedded peak using the minimum entropy principle; D. eROI-SIA resolution results; E. Validation of GC×GC separation.
[0018] Figure 4 A. eROI processing results of two raw data (30.8-31.2 minute segments); B. Rank plot; C. MSCC and IP-MSCC between matrices X and Y; D. Common components in matrices X and Y; E. eROI-AMWFA scanning results of the target region.
[0019] Figure 5 A. eROI processing of fingerprints in the 35.55-35.80 minute segment; B. Inputting the matrix into the MCR-ALS software; C. eROI-MCR-ALS parsing results.
[0020] Figure 6 A. GC-Orbitrap HRMS test of "Gui Yu Jin"; B. eROI compression; C. eROI-SIA parsing results; DGC×GC verification.
[0021] Figure 7 A. Non-targeted GC-HRMS fingerprint testing of “Green Tulip”; B. eROI compression and background subtraction. Detailed Implementation
[0022] The present invention will be further illustrated below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Experimental methods in the following embodiments, unless otherwise specified, are generally performed under conventional conditions or as recommended by the manufacturer. Unless otherwise stated, all percentages, ratios, proportions, or parts are by weight.
[0023] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as are familiar to those skilled in the art. Furthermore, any methods and materials similar to or equivalent to those described herein may be applied to the methods of this invention. The preferred embodiments and materials described herein are for illustrative purposes only.
[0024] Example 1: eROI Compression of Non-Targeted GC-HRMS Fingerprints of Biological Samples
[0025] like Figure 1As shown in (A), the developed eROI program consists of three steps. First, eROI uses NumPy to read the mzML file and iterates through each mass spectrum, extracting the retention time, mass-to-charge ratio (m / z), and corresponding peak intensity. Simultaneously, intensities less than 1000 are set to zero, and the m / z values for each retention time are reduced to low-resolution values, grouped, and their intensities are summed. Second, eROI stores the results in a dictionary, where the key is the retention time and the value is a dictionary containing the sum of peak intensities for each m / z range. eROI converts the results to a DataFrame format, sorts the column labels by m / z range, and writes the results to an Excel file. Third, eROI uses pandas to read the Excel file, extracts the data from the first column, groups it by the integer part of the retention time, interpolates within each integer part, and saves it to a new Excel file. Figure 1 As shown in (B), eROI can be converted into a mat file for easy calculation by MATLAB software.
[0026] exist Figure 2 In (A), a derivatized biological sample exhibited dense chromatographic peaks and significant peak overlap after GC-Orbitrap HRMS testing. Figure 2 In (B), the researchers converted the raw data into a matrix format using eROI, which allowed them to clearly observe the "background" and "co-flow" phenomena from the elution curves of each ion, highlighting the urgent need for chemometric methods to process the data.
[0027] Example 2: eROI-MCR-assisted acquisition of pure mass spectrometry data of sesquiterpene metabolites in the non-targeted GC-HRMS fingerprint of Cyperus rotundus.
[0028] like Figure 3 As shown in the total ion chromatogram (A), the GC separation of the (semi)volatile metabolite groups of the *Cyperus rotundus* sample was good, and the Orbitrap-HRMS detection had a wide linear range, capable of detecting trace components and not easily exhibiting signal saturation. Therefore, the peak area normalization method can effectively perform quantitative analysis of each sesquiterpene metabolite. Of course, non-targeted GC-Orbitrap-HRMS testing also generates matrix (or tensor) type data, which can be used to obtain a list of candidate sesquiterpene metabolites through mass spectrometry similarity retrieval, and finally the retention index is used to confirm the target components. However, some "co-current" peak clusters still exist, requiring further deconvolution using eROI-MCR.
[0029] eROI data structures can be deconvolved with "co-current" peak clusters using the classic MCR algorithm. However, depending on the type of "co-current" scenario, appropriate HELP, SIA, AMWFA, and MCR-ALS algorithms must be selected to accurately obtain the pure mass spectra of the corresponding peak clusters. In this embodiment, a detailed explanation is provided using the non-targeted GC-Orbitrap-HRMS fingerprint of the medicinal plant "Cyperus rotundus" as an example.
[0030] (1) Single matrix: eROI-SIA deconvolution has only two components of "embedded" peak clusters.
[0031] The "embedded" peak lacks a pure component region, making deconvolution unsuitable for methods or platforms like ROI-MCR, ADAP-GC 4.0, and MS-DIAL. In such cases, the eROI-SIA algorithm can be used. Figure 3 In our previous work, we demonstrated eROI-SIA deconvolution of GC-HRMS data. Taking the 33.75-33.90 min segment of the metabolome fingerprint of Cyperus rotundus essential oil as an example, the two components embedded in this peak cluster are difficult to observe from the total ion spectrum. However, the mass spectrometric difference between the two scan points on the left and right sides of this segment can indirectly indicate the existence of component co-flow. Unfortunately, researchers have found it difficult to obtain satisfactory results when using MS-DIAL and ADAP (MZmine2) to resolve this embedded peak cluster. This is because the principles of these two methods are similar to (non-)iterative multivariate resolution methods, and the deconvolution process depends on the extraction of pure component regions. To solve this problem, researchers can resort to the SIA method. Figure 3 As shown in (B), the original data is first compressed using eROI to reduce its dimensionality to an equal-row, equal-column matrix, and then preprocessed with smoothing and background subtraction. Subsequently, eROI-SIA is used to resolve the deconvolution of the two embedded peaks. Figure 3 In (C), a selective ion detection map was plotted using the minimum entropy principle. From the two edge lines passing through the origin, we can find two selective ions (m / z 96 and m / z 159) representing the peak cluster. Then, the algorithm extracts the concentration distribution from the selective ions and calculates the pure mass spectrum using the least squares method. Figure 3 The analysis results of (D1) and (D2) indicate that one is caryophyllene oxide and the other is β-Copaen-4α-ol. Figure 3 In (E), the researchers also conducted GC×GC (inverted column set) separation on the essential oil of Cyperus rotundus and found the peaks of Caryophyllene oxide and β-Copaen-4α-ol from the two-dimensional fingerprint, thus verifying the accuracy of the above algorithm.
[0032] (2) Two matrices: eROI-AMWFA deconvolution of "overlapping" peaks with large background interference.
[0033] For co-current peak clusters, researchers can also use the eROI-AMWFA method for analysis. The AMWFA method is suitable for comparing the MSCC of multiple peak clusters in pairwise sample fingerprints, obtaining elution curves and mass spectrometry data for a single common component. We will explain the steps using the 30.8-31.2 minute interval of the *Cyperus rotundus* metabolome data as an example. Figure 4 As shown in (A), the original matrices X and Y of the two Cyperus rotundus GC-Orbitrap HRMS fingerprints are obtained through eROI processing. Subsequently, in Figure 4 (B) In the rank plot, at least four compounds were observed to overlap or be embedded together, especially component "1," which was interfered with by multiple extraneous peaks and could not be deconvolved by MS-DIAL and ADAP (MZmine2). Therefore, we used the eROI-AMWFA algorithm to mine this segment of data. First, the X and Y matrices were subjected to singular value decomposition (SVD) to obtain the multi-component spectral correlation chromatography (MSCC) and back-projection multi-component spectral correlation chromatography (IP-MSCC) results between them. Figure 4 (C) indicates that the mass spectrometric characteristics of compounds in X are highly correlated with those of compounds in Y. While... Figure 4 In (D), it is also shown that X and Y have 4 common components. Then, a moving window search is performed between the X and Y matrices with a fixed window size, and the two orthogonal loading bases are E and F (E = [e1,e2,...,em] in X; F = [f1,f2,...,fn] in Y); m and n are the number of chemical components in matrices X and Y. Figure 4 In (E), after mutual scanning and calculation of the regions of X and Y each containing the "1" component, the shared component region, the spectral autocorrelation curve, and the mass spectrum of the "1" component (4,5-di-epi-aristolochene) can be observed. In addition, the pure mass spectra of 2 (Ylanga-2,4-diene) and 3 (γ-Muurolene) were also subsequently extracted.
[0034] (3) One or more matrices: eROI-MCR-ALS deconvolution of multiple "overlapping" peak clusters
[0035] Researchers can further combine eROI with the MCR-ALS method to apply it to deconvolution scenarios involving most overlapping peak clusters, improving both qualitative and quantitative analysis. Unlike AMWFA and HELP, MCR-ALS is an iterative multivariate resolution method with well-developed software and superior ability to resolve confluence peaks. Figure 5As shown in (A), the 35.55-35.80 segment of the Cyperus rotundus fingerprint shows an overlapping peak cluster containing three secondary metabolites and one interfering compound. Due to the absence of the pure component region, MS-DIAL has difficulty obtaining mass spectrometry information of the interfering component and component 2, and even more difficulty in resolving the quantitative curve. Through eROI processing, researchers can obtain an equal-row, equal-column matrix of this peak cluster in different samples, and further perform background subtraction and baseline correction. Figure 5 As shown in (B), these matrices are input into the MCR-ALS software and deconvolved under appropriate parameter settings. Its detailed working principle can be found in the references. Subsequently, in Figure 5 In (C), chromatographic and mass spectrometric data for three components and one interfering compound were successfully obtained. Component 1 was Cadalene, component 2 was Mustakone, and component 3 was an unknown compound. Furthermore, these resolved mass spectra and GC×GC-TOFMS fingerprints were validated, showing good agreement between fragment ions and relative intensities. More importantly, MCR-ALS provides near-perfect quantitative information. However, it also has several drawbacks: it is not suitable for embedded peaks; background subtraction of fingerprint peaks is required; it is sensitive to the number of components; and it exhibits rotational uncertainty.
[0036] In summary, the above-mentioned multi-dimensional resolution algorithms should be used according to different application scenarios.
[0037] Compared to previous ROI-MCR-ALS methods, eROI's advantage lies in not requiring multiple datasets to be aligned in the m / z column, and its ability to accurately read the m / z values from deconvolution mass spectra. Its disadvantage is the lack of precision in the m / z values, requiring users to manually add high-resolution values based on the original data.
[0038] Example 3: eROI-assisted acquisition of pure mass spectrometry data of sesquiterpene metabolites in the non-targeted GC-HRMS fingerprint of "Cinnamomum cassia"
[0039] Sesquiterpenes are a major component of the volatile oil of Curcuma longa, a medicinal herb, and have been proven to possess a wide range of biological activities. However, Curcuma longa exhibits a strong regional character, with the types and contents of sesquiterpenes often closely related to the species and origin. Among them, *Curcuma kwangsiensis* is one of four types of Curcuma longa medicinal herbs and is also a "genuine" herb of Guangxi Zhuang Autonomous Region, recorded in the 2020 edition of the Chinese Pharmacopoeia as the dried tuberous root of *Curcuma kwangsiensis*. However, the original source of *Curcuma kwangsiensis* cultivated in some areas of Guangxi Zhuang Autonomous Region has changed, and even *Curcuma zedoaria* and *Curcuma longa* from the same producing area do not belong to the same origin. This means that the phytochemical components and quality markers of this type of medicinal herb need to be re-examined. Therefore, conducting sesquiterpene component analysis on the volatile oil of mainstream *Curcuma kwangsiensis* medicinal herbs in Guangxi Zhuang Autonomous Region is an important prerequisite for the quality evaluation of traditional Chinese medicine.
[0040] To further understand the mainstream medicinal material "Cinnamon Turmeric," we collected samples from Luwu Town and Wuli Town in Lingshan County, Qinzhou City, Liangjiang Town in Wuming District, Nanning City, and Daqiao Town in Binyang County, all areas with significant production. Non-targeted GC-Orbitrap-HRMS analysis was performed. The total ion chromatogram (chromatographic vector) showed good GC separation of the "Cinnamon Turmeric" essential oil components, with most components achieving perfect separation. More importantly, GC-Orbitrap-MS exhibits a wide linear range, capable of detecting trace components in volatile oils, and is less prone to signal saturation. Therefore, the researchers used peak area normalization to calculate the relative percentage content of "Cinnamon Turmeric" essential oil components, enabling effective quantitative analysis. Furthermore, non-targeted GC-HRMS analysis can also generate matrix (or tensor) data, including both chromatographic and mass spectrometric information. Using TraceFinder... TM 4.1 After background subtraction using software, we obtained a clean high-resolution mass spectrum. Then, we obtained a list of candidate compounds through similarity search, and finally used the retention index to confirm the target component.
[0041] However, the "turmeric" fingerprint still contains clusters of confluence peaks. Figure 6 In our previous work, we explained the eROI-MCR of this data set. Two component-embedded peak clusters exist in the 35.9-36.1 min region of the "Cinnamomum cassia" fingerprint, which are difficult to observe from the total ion chromatogram. However, the mass spectrometric difference between the two scan points on either side of this region indirectly indicates the presence of component co-current. Unfortunately, researchers have struggled to obtain satisfactory results when using MS-DIAL and ADAP (MZmine2) to resolve these embedded peak clusters. This is because these two methods are similar in principle to (non-)iterative multivariate resolution methods, where the deconvolution process depends on the extraction of pure component regions. To solve this problem, researchers can resort to parameter-free SIA methods. Figure 6 As shown in (A), the original data is first compressed using eROI, reducing its dimensionality to an equal-row, equal-column matrix, and then preprocessed with smoothing and background subtraction. Subsequently, eROI-SIA is used as two embedded peaks for deconvolution. Figure 6 In (B), a selective ion detection map was plotted using the minimum entropy principle. Two selective ion peak clusters (m / z 108 and m / z 135) were identified from the two edge lines passing through the origin. The algorithm then extracted the concentration distribution from the selective ions and calculated the pure mass spectrum using the least squares method. Figure 6 As shown in (C1) and (C2), the identification results indicate that one is furandiene and the other is gimazone. Researchers also conducted GC×GC separation on the "Cinnamomum cassia" essential oil and extracted the results from the two-dimensional fingerprint. Figure 6 (D) The peaks for furandiene and gemmaconone were found, verifying the accuracy of the above algorithm.
[0042] Example 4: eROI-assisted acquisition of pure mass spectrometry data of sesquiterpene metabolites in the non-targeted GC-HRMS fingerprint of "Green Silk Tulip".
[0043] Researchers also tested the "Green Silk Turmeric" essential oil from Sichuan. For example... Figure 7 As shown in (A), the GC separation of the "Green Turmeric" essential oil components was good, with most components achieving perfect separation. However, this fingerprint segment had a low signal-to-noise ratio, significant noise influence, and the presence of confluence peak clusters. Figure 7 In (B), the original fingerprint is first compressed by eROI, reduced to an equal-row, equal-column matrix, and preprocessed by smoothing and background subtraction. Finally, it is deconvolved to obtain Curcumenol.
[0044] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of the substantive technical content of the present invention. The substantive technical content of the present invention is broadly defined within the scope of the claims. Any technical entity or method completed by others that is completely identical to or an equivalent modification of the claims is considered to be covered within the scope of the claims.
Claims
1. A method for isometric region of interest (eROI) compression of gas chromatography-high resolution mass spectrometry fingerprint data, characterized in that: The normalized fingerprint is derived from GC-HRMS data of complex samples and is also applicable to LC-HRMS data. It includes three main steps: EICs extraction, column equalization, and row equalization. The steps for eROI compression are as follows: (1) Sample Testing: ① Obtain the characteristic component extract of the test sample using standardized procedures. Its composition and structure should truly represent the components within the sample. ② If the test sample is a biological sample, a quantitative sample should be weighed, derivatized, and then subjected to GC-Orbitrap HRMS testing. Methodological investigation should be conducted to ensure that the GC peak separation is as complete as possible, with appropriate intensity and reasonable distribution. ③ If the component extract is a plant or food essential oil or other easily (semi) volatile liquid, the test sample should be diluted with an equal volume of n-hexane, injected appropriately, and the chromatographic column optimized to ensure that the GC peak separation is as complete as possible, with appropriate intensity and reasonable distribution. The above test data should have good reproducibility and strong characteristic properties. (2) EICs Extraction: ① Convert the raw data into a common format file using software. ② eROI reads the common file (e.g., mzML) and iterates through each mass spectrum to extract retention time, mass-to-charge ratio (m / z), and corresponding peak intensity. ③ Threshold the peak intensities, setting intensities below the threshold to zero, and reduce the m / z value at each retention time to a low-resolution value, grouping them and summing the intensities. (2) Column Equalization: ① eROI stores the results in a dictionary, where the key is the retention time and the value is a dictionary containing the sum of peak intensities corresponding to each m / z range. ② eROI sorts the column labels according to the m / z range, so that each m / z falls into a specific position, and writes the results to an Excel file. (3) Row alignment: ① eROI reads the Excel file above and extracts the data in the first column. ② The data is grouped according to the integer part of the retention time, and interpolation is performed in each integer part. The entire data is traversed to complete the row alignment interpolation, and the inserted data is saved to a new Excel file. (4) eROI-MCR: Convert the Excel file processed above into a mat file (equal row and column matrix), and combine it with the classic "multivariate resolution (MCR)" or "multidimensional correction" algorithm for deconvolution of "common flow peaks" and obtain accurate qualitative or quantitative data.
2. The eROI compression method according to claim 1, characterized in that, Combining column equal spacing and row equal spacing.
3. The eROI compression method according to claim 1, characterized in that, GC-HRMS or LC-HRMS fingerprints are suitable for testing extremely complex samples under standard experimental conditions. They can be used for both pure spectral deconvolution for qualitative analysis and quantitative analysis.
4. The eROI compression method according to claim 1, characterized in that, The column equidistancy refers to ensuring that the peak intensity corresponding to each m / z range falls into a specific position so that the m / z value can be quickly read in the subsequent mass (wave) spectrum analysis; at the same time, it also makes it so that the fingerprints of different samples do not need to be aligned with columns of multiple sets of data after compression.
5. The eROI compression method according to claim 1, characterized in that, The row equalization involves grouping the data according to the integer part of the retention time and interpolating within each integer part, so that different fingerprints do not need to be aligned with multiple groups of data after compression.
6. The eROI compression method according to claim 1, characterized in that, The .mat file processed by eROI can be combined with classic MCR or "multidimensional correction" algorithms to perform deconvolution. This includes, but is not limited to, the following methods: Selective Ion Analysis (SIA), Alternating Moving Window Factor Analysis (AMWFA), Multivariate Curve-Resolved Alternating Least Squares (MCR-ALS), Alternating Trilinear Decomposition (ATLD), etc.
