UHPLC-HRMS profile mode data centroid conversion method

Abstract

The invention discloses a UHPLC-HRMS profile mode data centroid conversion method, which comprises the following steps: acquiring profile mode data based on UHPLC-HRMS, and extracting a local maximum value vector based on each piece of mass spectrum data in the profile mode data; obtaining an outlier vector and a candidate peak vector through outlier detection based on the local maximum vector; estimating instrument noise based on the outlier vector and the candidate peak vector; and centroid ion extraction is carried out based on the instrument noise and the candidate peak vector, and UHPLC-HRMS profile mode data centroid conversion is completed. According to the method, information of metabolites in each mass spectrum can be effectively and clearly separated from information of background noise, the risk of compound information loss is reduced, and effective conversion of UHPLC-HRMS profile mode data is realized.

Description

Technical Field

[0001] The invention belongs to the technical field of ultra-high performance liquid chromatography-high resolution mass spectrometry data analysis, and specifically relates to a method for centroiding conversion of UHPLC-HRMS profile mode data. Background Art

[0002] Ultra-High Performance Liquid Chromatography-High Resolution Mass Spectrometry (UHPLC-HRMS) is widely used in omics research, including metabolomics and lipidomics. In routine analysis, most vendors' instruments, such as Agilent, Waters, ThermoFisher, etc., support centroid and profile (or Continues) modes for UHPLC-HRMS data acquisition. Compared with the centroid mode, one advantage of the profile mode is that it can retain all the information in the mass spectrum to the maximum extent, so a fairly large data file can be provided for each sample analyzed. For researchers, it is by no means an easy task to analyze the data in the UHPLC-HRMS profile mode efficiently and with high quality.

[0003] There is no obvious difference between UHPLC-HRMS profile mode data analysis and centroid mode data analysis, except that a process of data centroid conversion is required. In short, the workflow of profile mode data analysis includes centroid conversion, construction of extracted ion chromatograms (EICs), chromatographic peak extraction and annotation, and chromatographic peak alignment and registration. Current UHPLC-HRMS data analysis tools can be briefly divided into two categories based on the centroid conversion achieved: i) centroid conversion is completed with the help of third-party software such as MSConvert; ii) the data analysis tool itself supports centroid conversion.

[0004] The main purpose of centroid conversion is to clearly separate the information of metabolites within each mass spectrum from the information of background noise. Published data analysis tools have adopted various strategies to achieve this goal. For example, the most effective centroid conversion can be performed by picking out the maximum value of each mass spectral peak. In this case, centroid conversion can be performed quickly during data acquisition and is therefore adopted by vendor software such as Agilent's MassHunter. XCMS further adopts a smoothing strategy to enhance the real peak signal in the original mass spectrum before mass spectral peak extraction. In MS-DIAL, a mass spectral peak extraction algorithm based on differential calculation and noise estimation is used to extract the peaks in each mass spectrum to perform centroid conversion. Another more complex strategy can be found in MSConvert, which uses wavelet analysis to detect mass spectral peaks in mass spectra for centroid conversion. Although various strategies have been developed for current data analysis tools, it seems that none of them have optimized the impact of centroid conversion on downstream data analysis steps, such as EIC construction, chromatographic peak extraction, etc. In summary, it is still meaningful to develop high-quality data analysis tools to effectively process UHPLC-HRMS profile mode data sets in complex sample analysis. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention provides a method for centroiding UHPLC-HRMS profile mode data, which can effectively realize the centroiding conversion of UHPLC-HRMS profile mode data and improve the impact on downstream data analysis steps, such as EIC construction and chromatographic peak extraction.

[0006] To achieve the above object, the present invention provides a method for centroiding data of a UHPLC-HRMS profile mode, comprising the following steps:

[0007] (1) obtaining profile mode data based on UHPLC-HRMS acquisition, and extracting a local maximum value vector based on each mass spectrum data in the profile mode data: convolving and smoothing the obtained mass spectrum with a Gaussian function to obtain a convolution mass spectrum signal; extracting a local maximum value based on the convolution mass spectrum signal to obtain a local maximum value vector;

[0008] (2) Based on the local maximum vector, an outlier vector and a candidate peak vector are obtained by outlier detection: outlier detection is performed using a moving window strategy, and the signal-to-noise ratio of the elements in the local maximum vector at the center of the window is calculated using the first-order derivative and robust statistics method, and the elements with a signal-to-noise ratio exceeding a first threshold are regarded as outlier elements, and the rest are candidate peak elements, and the above steps are repeated until the window traverses the convolution mass spectrometry signal to obtain the outlier vector and the candidate peak vector;

[0009] (3) estimating the instrument noise based on the outlier vector and the candidate peak vector: using the original position in the outlier vector to search in the original mass spectrum and mark it as a fixed point, using a linear interpolation strategy to estimate the missing value between two fixed points, and constructing the instrument noise at each point in the profile mode mass spectrum;

[0010] (4) Centroid ion extraction is performed based on the instrument noise and candidate peak vectors to complete the centroid conversion of UHPLC-HRMS profile mode data: the ion abundance of the elements in the candidate peak vector is divided by the corresponding estimated instrument noise to calculate the signal-to-noise ratio of each element; the elements in the candidate peak vector whose signal-to-noise ratio is higher than the second threshold are retained, and the m / z value and ion abundance correction are performed for the m / z difference between two adjacent elements within the tolerance range, and finally the centroid ion extraction is completed to achieve the centroid conversion.

[0011] Furthermore, in the step (1), the Gaussian function used for the convolution smoothing is a 5-point Gaussian function.

[0012] Furthermore, in step (1), the method for extracting the local maximum value is: for any convolution mass spectrometry signal s, by s i-1 >s i &s i i+1 Extract the local maximum value; where s i is the ith data point in the mass spectrum, s i-1 and i+1 They are the i-1th and i+1th data points respectively.

[0013] Furthermore, in the step (2), in the moving window strategy, the window size is 30.

[0014] Furthermore, in the step (2), in the outlier detection using the moving window strategy, the first threshold of the signal-to-noise ratio is 2.5.

[0015] Furthermore, in the step (4), in the method of extracting centroid ions using the instrument noise and the candidate peak vector, the second threshold of the signal-to-noise ratio is 1.5, and the m / z tolerance of two adjacent elements is 0.015.

[0016] Compared with the prior art, the present invention has the following advantages and technical effects:

[0017] ​The present invention provides a UHPLC-HRMS profile mode data centroid conversion method, which is based on the maximum value centroid transformation (Maximum Value-based centroiding Transformation, MVCT) strategy, and its basic principle is that the ion corresponding to the compound in the profile mode mass spectrum should be the local maximum value. The method first convolutes and smoothes the obtained mass spectrum with a Gaussian function, and then extracts the maximum value to obtain a local maximum value vector; outlier detection is performed using a moving window strategy, and the signal-to-noise ratio is calculated using a first-order derivative and a robust statistical method to obtain the outlier vector and the candidate peak vector; according to the outlier vector, the linear interpolation strategy is used to estimate the mass spectrometer noise; the ions of the elements in the candidate peak vector are screened according to the signal-to-noise ratio, and the m / z value and ion abundance correction are performed for the m / z difference between two adjacent elements within the tolerance range, and finally the extraction of the centroid ion is completed to achieve centroid conversion. The method converts the centroid data to effectively separate the information of the metabolites in each mass spectrum from the information of the background noise, and reduces the risk of compound information loss. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0019] Figure 1 This is a schematic diagram of a method for centroid conversion of UHPLC-HRMS profile mode data according to an embodiment of the present invention;

[0020] Figure 2 is an overall workflow diagram of outlier detection in an embodiment of the present invention;

[0021] Figure 3 : is a schematic diagram of a method for centroiding conversion of UHPLC-HRMS profile mode data in an embodiment of the present invention, wherein: Figure 3 a in is the original profile signal of the mass spectrum, Figure 3 b in the figure is the mass spectrum after analysis, where the maximum value in x is represented by a point. Figure 3 The c in the graph is the x in the selected spectrum. peak , Figure 3 The d in the graph is the x in the selected spectrum. noise , Figure 3 where e is the estimated instrument noise, Figure 3 Where f is x peak The signal-to-noise ratio corresponding to the elements in is, Figure 3 The g in is the final centroid result. Figure 3The h in is the centroided signal in the mass spectrum space;

[0022] Figure 4 FIG. 1 is a schematic diagram of a brief comparison between the present invention and MSConvert in an embodiment of the present invention, wherein: Figure 4 a in the figure is the comparison of EIC number, MS 1 Chromatographic peaks, MS with MS / MS spectra 1 Chromatographic peaks and MS with MF>0.7 1 Chromatographic peak, Figure 4 b in the figure is an example of mass spectrometry peak extraction;

[0023] Figure 5 Schematic diagram of mass spectrum peak extraction under different parameter settings of MSConvert in an embodiment of the present invention, wherein: Figure 5 a in the figure is a schematic diagram of the mass spectrum peak extraction results using the CWT method (Min SNR = 0.1, Min peak spacing = 0.01) by MSConvert. Figure 5 b in the figure is a schematic diagram of the mass spectrum peak extraction results using the CWT method (Min SNR = 0.01, Min peak spacing = 0.1) by MSConvert. Figure 5 c in the figure is a schematic diagram of the mass spectrum peak extraction results using the CWT method (Min SNR = 0.01, Minpeak spacing = 0.01) by MSConvert. Figure 5 d in the figure is a schematic diagram of the mass spectrometry peak extraction results using the Vendor method by MSConvert;

[0024] Figure 6 Schematic diagram of the centroid mass spectrum with a retention time of 27.023 min obtained by the present invention and MSConvert in the embodiment of the present invention, wherein: Figure 6 a in the figure is a schematic diagram of the original profile mass spectrum with a retention time of 27.023 min. Figure 6 b in the figure is a schematic diagram of the centroid mass spectrum obtained by the present invention, Figure 6 c in the figure is a schematic diagram of the centroid mass spectrum obtained by MSConvert using the Vendor method. Figure 6 The d in the figure is a schematic diagram of the centroid mass spectrum obtained by MSConvert using the CWT method (Min SNR = 0.1, Min peak spacing = 0.1). Figure 6e in the figure is a schematic diagram of the centroid mass spectrum obtained by MSConvert using the CWT method (Min SNR = 0.1, Min peak spacing = 0.01). Figure 6 f in the figure is a schematic diagram of the centroid mass spectrum obtained by MSConvert using the CWT method (Min SNR = 0.01, Min peak spacing = 0.1). Figure 6 g in the figure is a schematic diagram of the centroid mass spectrum obtained by MSConvert using the CWT method (Min SNR = 0.01, Min peak spacing = 0.01);

[0025] Figure 7 Schematic diagram of the mass spectrum peak extraction process in CWT in an embodiment of the present invention, wherein: Figure 7 a in the figure is a schematic diagram of the original profile spectrum. Figure 7 The b in the figure is a schematic diagram of wavelet coefficients at different wavelet scales. Figure 7 The c in FIG. is a schematic diagram of the contour map of the signal in the wavelet scale space. DETAILED DESCRIPTION

[0026] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0027] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0028] Example

[0029] The following four plant matrices, licorice, chrysanthemum, honeysuckle and tobacco, were used as samples and analyzed using Waters ACQUITY UPLC-AB SCIEX TripleTOF TM UHPLC-HRMS analysis was performed using 5600plus MS.

[0030] First, the profile mode data required by the present invention is acquired based on UHPLC-HRMS. In this embodiment, the following steps are specifically included:

[0031] (1) Preparation of quality control samples: The preparation method of quality control samples is based on licorice. Weigh the powder of each licorice sample by equal weight and mix them evenly to prepare the licorice QC sample. The QC samples of chrysanthemum, honeysuckle and tobacco leaves are prepared in the same way. All prepared QC samples are stored in a -40℃ refrigerator for analysis and used for the analysis of complex plant matrices.

[0032] (2) Sample pretreatment: For four plant matrices, including licorice, chrysanthemum, honeysuckle and tobacco leaves, approximately 20 mg of powder was weighed into an Eppendorf tube (2 mL specification). Approximately 1.5 mL of methanol (CH 3 OH / H 2 O=7 / 3) extract for metabolite extraction. After vortexing for 2 min, the centrifuge tube was placed at room temperature for ultrasonic treatment for 30 min. Then, the centrifuge tube was centrifuged at 13000 rpm for 10 min, and the supernatant was transferred to a chromatographic bottle for UHPLC-HRMS analysis.

[0033] (3)Use Waters ACQUITY UPLC-AB SCIEX TripleTOF TM 5600plus MS analysis of licorice samples:

[0034] The chromatographic conditions for analyzing the licorice sample were as follows: Waters ACQUITY UPLC BEH C18 column (100 mm × 2.1 mm, 1.7 μm particle size) was used for separation. The column temperature was set at 35°C and the flow rate was 0.2 mL / min. The mobile phases A and B were 0.1% formic acid in water and 0.1% formic acid in acetonitrile, respectively. The elution gradient is shown in Table 1. The elution program was terminated within 36 min, followed by 2 min of column equilibration. The sample injection volume was 4 μL.

[0035] Table 1

[0036]

[0037] The mass spectrometry conditions for analyzing licorice samples were as follows: ionization of metabolites and MS detection were achieved by electrospray ionization (ESI). MS / MS spectra of dependency (IDA) data of all licorice samples were acquired using positive ion mode. The mass spectrometry conditions included MS (TOF-MS) and MS / MS (Product Ion) detection parameters. The detailed parameter settings are as follows:

[0038] TOF-MS parameters: mass scanning range 100~1000Da; accumulation time 0.18s; ion source gas (GS1) flow rate 50mL / min; ion source gas (GS2) flow rate 50mL / min; curtain gas (CUR) 30mL / min; temperature (TEM) 500V; Ion Sparay Voltage Floating (ISVF) 5500V; declustering voltage (DP) 80V; collision energy (CE) 10V.

[0039] Product Ion parameters: mass scanning range 50-1000Da; accumulation time 0.03s; cycle time 0.35s; ion source gas (GS1) flow rate 50mL / min; ion source gas (GS2) flow rate 50mL / min; curtain gas (CUR) 30mL / min; temperature (TEM) 500V; Ion Sparay Voltage Floating (ISVF) 5500V; declustering voltage (DP) 80V; collision energy (CE) 35V; collision energy spread (CES) 15V; ion release delay (IRD) 67; ion release width (IRW) 25; mass tolerance 100mDa; maximum number of candidate ions monitored per cycle 5; exclude the former target ions of 5s.

[0040] (4)Use Waters ACQUITY UPLC-AB SCIEX TripleTOF TM 5600plus MS analysis of chrysanthemum samples:

[0041] The chromatographic conditions for analyzing the chrysanthemum sample were as follows: a Waters ACQUITY UPLC BEH C18 column (100 mm × 2.1 mm, 1.7 μm particle size) was used for separation. The column temperature was set at 35 °C and the flow rate was 0.2 mL / min. The mobile phases A and B were 0.1% formic acid in water and 0.1% formic acid in acetonitrile, respectively. The elution gradient is shown in Table 2. The elution program was terminated within 40 min, followed by 3 min of column equilibration. The sample injection volume was 4 μL.

[0042] Table 2

[0043]

[0044] The mass spectrometry conditions for analyzing the chrysanthemum sample are as follows: the mass spectrometry instrument parameters selected for the chrysanthemum sample analysis are consistent with the mass spectrometry parameters selected for the licorice sample analysis. For details, please refer to the UHPLC-HRMS instrument analysis conditions for licorice samples.

[0045] (5)Use Waters ACQUITY UPLC-AB SCIEX TripleTOFTM 5600plus MS analysis of honeysuckle samples:

[0046] The chromatographic conditions for analyzing the honeysuckle sample were as follows: a Waters T3 column (100 mm × 2.1 mm, 1.7 μm particle size) was used for separation. The column temperature was set at 35 °C and the flow rate was 0.2 mL / min. The mobile phases A and B were 0.1% formic acid in water and 0.1% formic acid in acetonitrile, respectively. The elution gradient is shown in Table 3. The elution program was terminated within 64 min, followed by 5 min of column equilibration. The sample injection volume was 4 μL.

[0047] Table 3

[0048]

[0049] The mass spectrometry conditions for analyzing the honeysuckle sample are as follows: the mass spectrometry instrument parameters selected for the honeysuckle sample analysis are consistent with the mass spectrometry parameters selected for the licorice sample analysis. For details, please refer to the UHPLC-HRMS instrument analysis conditions for licorice samples.

[0050] (6)Use Waters ACQUITY UPLC-AB SCIEX TripleTOF TM 5600plus MS analysis of tobacco samples:

[0051] The chromatographic conditions for analyzing tobacco samples were as follows: a Waters T3 column (100 mm × 2.1 mm, 1.7 μm particle size) was used for separation. The column temperature was set at 35°C and the flow rate was 0.2 mL / min. The mobile phases A and B were 0.1% formic acid in water and 0.1% formic acid in acetonitrile, respectively. The elution gradient is shown in Table 4. The elution program was terminated within 30 min, followed by 3 min of column equilibration. The sample injection volume was 4 μL.

[0052] Table 4

[0053]

[0054] The mass spectrometry conditions for analyzing tobacco samples are as follows: the mass spectrometry instrument parameters selected for the analysis of honeysuckle samples are consistent with the mass spectrometry parameters selected for the analysis of licorice samples. For details, please refer to the UHPLC-HRMS instrument analysis conditions for licorice samples.

[0055] Secondly, the UHPLC-HRMS profile mode data centroid conversion method, such as Figure 1 As shown, the specific steps include:

[0056] S1. acquiring profile mode data based on UHPLC-HRMS, and extracting a local maximum value vector based on each mass spectrum data in the profile mode data;

[0057] In this embodiment, the local maximum value of each mass spectrum data is extracted:

[0058] The obtained mass spectrum s is convolved with a Gaussian function. Then, according to the obtained convolution mass spectrum signal, s i-1 >s i &s i i+1 Extract the local maximum in the mass spectrum. Where s i is the ith data point in the mass spectrum, s i-1 and i+1 They are the i-1th and i+1th data points respectively. Thus, a new vector x containing only the extracted local maxima can be obtained, and its elements are arranged in ascending order according to their original positions.

[0059] Among them, the Gaussian function used for convolution smoothing is a 5-point Gaussian function.

[0060] S2. Based on the local maximum vector, obtaining an outlier vector and a candidate peak vector by outlier detection;

[0061] In this embodiment, outlier vector and candidate peak vector extraction: an outlier detection step is performed on the local maximum vector x containing the noise from the instrument to distinguish the elements corresponding to the instrument background noise and the compound. Based on the basic fact that the elements from the compound should be much larger than the noise, if an element in x is much larger than its neighbor, it will be marked as an outlier. In this embodiment, a moving window strategy is used to find outliers. The elements at the center of the window will be verified by using robust statistics. At the same time, the first-order derivative of x is also used to check for outliers. Figure 2 The detailed process of local maximum extraction is shown. In the moving window strategy (left part), by S / N = max(|(x i -median(x w )) / σ|,|(x i -x i-1 ) / σ|) calculates the signal-to-noise ratio of the window center point, where x i is the center point of the moving window, x w is an element in the window vector. σ=1.483*median|x j -median(x w )|(j=1,…,J). σ is the estimated noise in the window; the constant 1.483 is a correction coefficient to make σ belong to the normal distribution. When the signal-to-noise ratio is greater than 2.5, x i ​will be considered as an outlier, corresponding to a 99% confidence level under the normal distribution. Regarding the strategy based on the first-order derivative (right part), the noise level of the vector x is given by Noise = 1.483*median|dx j -median(dx)|(j=1,…,N) is estimated, where dx is the first-order derivative of x; dx j is the jth data point in dx; N is the number of data points in dx. j The outlier vector is regarded as an outlier element, and the rest are candidate peak elements. The above steps are repeated until the window traverses the convolution mass spectrometry signal, and finally the outlier vector and candidate peak vector are obtained.

[0062] Among them, in the moving window strategy used, the window size is 30 and the signal-to-noise ratio threshold is 2.5.

[0063] S3. Estimating instrument noise based on the outlier vector and the candidate peak vector;

[0064] In this embodiment, the instrument noise estimation is: the outliers in x are temporarily marked as candidate peak vectors (x peak ), the rest of x is considered as the instrument noise vector (x noise ), and according to x noise Automatically estimate instrument noise. By using the outlier vector x noise The original positions of the elements in are retrieved in the original mass spectrum and marked as fixed points. Then, a linear interpolation strategy is used to estimate the missing values ​​between two consecutive fixed points to construct the instrument noise s at each point in the profile mode mass spectrum. noise .

[0065] S4. Perform centroid ion extraction based on the instrument noise and candidate peak vectors to complete the centroid conversion of UHPLC-HRMS profile mode data.

[0066] In this example, centroid ion extraction is performed by converting the candidate peak vector x peak The ion abundance of each element in x is divided by the corresponding estimated instrument noise to calculate peak The signal-to-noise ratio of each element in . If the signal-to-noise ratio is higher than the threshold x peak The elements in will be retained, and the m / z values ​​of the adjacent elements whose m / z difference is within the tolerance range will be passed. cen =(Int 1 ×m / z 1 +Int 2 ×m / z 2 ) / (Int 1 +Int 2 ) for m / z value and ion abundance correction, where Int 1 and Int2 are the ion intensities of the first and second ions, respectively. m / z 1 and m / z 2 is a high-precision m / z value that is within the preset m / z tolerance. cen is the high-precision m / z value of the merged ion. 1 and Int 2 The larger value of will be used as the ion intensity of the merged ion, that is, Int cen =max(Int 1 ,Int 2 ).

[0067] The signal-to-noise ratio threshold was set to 1.5, and the m / z tolerance of two adjacent elements was 0.015.

[0068] Some of the analytical data results of the present invention are as follows:

[0069] The working principle of the present invention is further described. Figure 3 As shown, Figure 3 Three profile mode mass spectra are provided in a, which were obtained at the 4500th, 4506th and 12827th scans, respectively. Figure 3 All local maxima extracted in the selected region of the 4500th spectra s are depicted in b. With the help of the developed outlier detection step, these local maxima are divided into two groups, x peak (See Figure 3 c) and x in noise (See Figure 3 d in). Then according to x noise , the instrument noise is adaptively estimated, such as Figure 3 e) in the figure. peak The signal-to-noise ratio of all elements in , and remove the elements with a signal-to-noise ratio lower than 1.5, such as Figure 3 After merging ions with m / z values ​​less than the preset m / z tolerance (0.015), a mass spectrum after mass center conversion can be obtained. Figure 3 The centroid conversion process will be performed independently on each profile mode mass spectrum to obtain the centroid mode mass spectrum after centroid conversion (see Figure 3 h) in.

[0070] A comprehensive study of the centroid conversion was performed using four food matrices from an untargeted dataset and compared with MSConvert. Figure 4As shown in Figure 2, MSConvert focuses on file format conversion and cannot provide other programs in UHPLC-HRMS data analysis. Therefore, in this invention, the centroid-converted file was imported into AntDAS-Profiler to provide a fair comparison. First, the number of EICs constructed and the number of detected MS 1 The performance is evaluated by the number of chromatographic peaks. Figure 4 As shown in a in the figure, it can be seen that MSConvert obtains the most EIC numbers in all four food matrices. However, the extracted chromatographic peaks clearly show that the present invention (AntDAS-Profiler in English) provides the highest number of EICs in all data sets. 1 The chromatographic peaks are much more than those of MSConvert. Since a method cannot be evaluated only by the extracted chromatographic peaks, the present invention also includes MS with MS / MS spectra. 1 The chromatographic peaks were compared because they may be more valuable for omics studies. Figure 4 It can be found in a that AntDAS-Profiler obtains more numbers in all food matrices than MSCovnert. In addition, this study further imported the MS / MS spectra of the two methods into a public compound library for compound identification, which can be downloaded from http: / / prime.psc.riken.jp / compms / msdial / main.html#MSP. The Match Factor (MF) is used to measure the quality of the MS / MS spectrum. It is calculated based on the dot product between the constructed MS / MS spectrum and the reference in the library, and the threshold of MF is set to 0.7. In this case, AntDAS-Profiler obtains more chromatographic peaks with MF>0.7 in all four matrices than MSConvert. In order to further clarify the difference between AntDAS-Profiler and MSConvert in centroid conversion, this embodiment is illustrated by an example, such as Figure 4 The original mass spectrometry signal is shown in Figure 4From the left part of b in the figure, it can be seen that there are two obvious regions of interest (ROI). The center of mass conversion result of MSConvert shows that only one ROI corresponding to m / z 425.2303 can be retained. Due to the low ion intensity, the ROI of m / z 425.1960 was filtered. In contrast, with the help of AntDAS-Profiler, two ROIs (m / z 425.2081, 425.2290) can be satisfactorily retained after the center of mass conversion. In addition, several ions located in the low intensity area can be detected. Then, the present invention optimizes the parameters of the center of mass conversion of the CWT strategy in MSConvert, and the results are shown in the attached figure. Figure 5 As shown; among them, Figure 5 a in the figure is a schematic diagram of the mass spectrum peak extraction results using the CWT method (Min SNR = 0.1, Min peak spacing = 0.01) by MSConvert. Figure 5 b in the figure is a schematic diagram of the mass spectrum peak extraction results using the CWT method (Min SNR = 0.01, Min peak spacing = 0.1) by MSConvert. Figure 5 c in the figure is a schematic diagram of the mass spectrum peak extraction results using the CWT method (Min SNR = 0.01, Min peakspacing = 0.01) by MSConvert. Figure 5 Figure d is a schematic diagram of the results of MSConvert using the Vendor method for mass spectrometry peak extraction. Unfortunately, none of them could successfully extract the EIC of m / z 425.2081. Figure 5 The results in b show that MSConvert may face problems in identifying ions without baseline separation. A very likely reason is the CWT-based centroid conversion strategy adopted in MSConvert.

[0071] In the attached Figure 6 A brief comparison of the centroid algorithms in AntDAS-Profiler and MSConvert is shown in Figure 6 a in the figure is a schematic diagram of the original profile mass spectrum with a retention time of 27.023 min. Figure 6 b in the figure is a schematic diagram of the centroid mass spectrum obtained by the present invention, Figure 6 c in the figure is a schematic diagram of the centroid mass spectrum obtained by MSConvert using the Vendor method. Figure 6 The d in the figure is a schematic diagram of the centroid mass spectrum obtained by MSConvert using the CWT method (Min SNR = 0.1, Min peak spacing = 0.1). Figure 6 e in the figure is a schematic diagram of the centroid mass spectrum obtained by MSConvert using the CWT method (Min SNR = 0.1, Min peakspacing = 0.01). Figure 6 f in the figure is a schematic diagram of the centroid mass spectrum obtained by MSConvert using the CWT method (MinSNR = 0.01, Min peak spacing = 0.1). Figure 6 The g in the figure is a schematic diagram of the centroid mass spectrum obtained by MSConvert using the CWT method (Min SNR=0.01, Min peak spacing=0.01). Figure 6 The a in Figure shows a raw mass spectrum with a retention time of 27.023 min. AntDAS-Profiler obtained 104 ions (see Figure 6 b). MSConvert only obtained 17 ions in the Vendor method (see Figure 6 c). By using the CWT method, MSConvert can obtain different ion counts under various parameters. When focusing on the ions around m / z 425.22, Figure 6 In the insert of a in Figure 2, two mass spectral peaks with no baseline separation can be found. Using AntDAS-Profiler, two ions can be successfully extracted. In contrast, only one ion can be obtained using MSConvert.

[0072] In order to clarify the reason, the present invention further conducts an in-depth study on the centroid conversion method based on CWT. The results are shown in the attached Figure 7 middle. Figure 7 The a in Figure 1 provides the original Profile mass spectrum recorded, where it can be seen that the two peaks are not baseline separated. The mother wavelet "Mexican-Hat" is used, and the signals at different wavelet scales are shown in Figure 7 In b, in addition, Figure 7 Figure c shows the contour map of the signal in the wavelet scale space. Obviously, the contour lines show that only one mass spectrum peak corresponding to the ion m / z 425.2304 can be extracted. Similar results can be obtained in the study of various mother wavelet functions. The above results show that the developed centroid conversion strategy is better than the strategy in MSConvert.

[0073] The above is only a preferred specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for centroiding UHPLC-HRMS profile data, characterized in that: The following steps are involved: (1) obtaining profile mode data based on UHPLC-HRMS acquisition, and extracting a local maximum value vector based on each mass spectrum data in the profile mode data: convolving and smoothing the obtained mass spectrum with a Gaussian function to obtain a convolution mass spectrum signal; extracting a local maximum value based on the convolution mass spectrum signal to obtain a local maximum value vector; (2) Based on the local maximum vector, an outlier vector and a candidate peak vector are obtained by outlier detection: outlier detection is performed using a moving window strategy, and the signal-to-noise ratio of the elements in the local maximum vector at the center of the window is calculated using the first-order derivative and robust statistics method, and the elements with a signal-to-noise ratio exceeding a first threshold are regarded as outlier elements, and the rest are candidate peak elements, and the above steps are repeated until the window traverses the convolution mass spectrometry signal to obtain the outlier vector and the candidate peak vector; (3) estimating the instrument noise based on the outlier vector and the candidate peak vector: using the original position in the outlier vector to search in the original mass spectrum and mark it as a fixed point, using a linear interpolation strategy to estimate the missing value between two fixed points, and constructing the instrument noise at each point in the profile mode mass spectrum; (4) Centroid ion extraction is performed based on the instrument noise and the candidate peak vector to complete the centroid conversion of the UHPLC-HRMS profile mode data: the ion abundance of the elements in the candidate peak vector is divided by the corresponding estimated instrument noise to calculate the signal-to-noise ratio of each element; the elements in the candidate peak vector whose signal-to-noise ratio is higher than the second threshold are retained, and the m / z value and ion abundance correction are performed for the m / z difference between two adjacent elements within the tolerance range, and finally the centroid ion extraction is completed to achieve the centroid conversion.

2. The UHPLC-HRMS profile mode data centroid conversion method according to claim 1, characterized in that: In the step (1), the Gaussian function used for the convolution smoothing is a 5-point Gaussian function.

3. The UHPLC-HRMS profile mode data centroid conversion method according to claim 1, characterized in that: In the step (1), the method for extracting the local maximum value is: for any convolution mass spectrometry signal s, by s i-1 >s i &s i i+1 Extract the local maximum value; where s i is the ith data point in the mass spectrum, s i-1 and i+1 They are the i-1th and i+1th data points respectively.​ 4. The UHPLC-HRMS profile mode data centroid conversion method according to claim 1, characterized in that: In the step (2), in the moving window strategy, the window size is 30.

5. The UHPLC-HRMS profile mode data centroid conversion method according to claim 1, characterized in that: In the step (2), in the outlier detection using the moving window strategy, the first threshold of the signal-to-noise ratio is 2.

5.

6. The UHPLC-HRMS profile mode data centroid conversion method according to claim 1, characterized in that: In the step (4), in the method of extracting centroid ions using the instrument noise and candidate peak vectors, the second threshold of the signal-to-noise ratio is 1.5, and the m / z tolerance of two adjacent elements is 0.015.

Images