Methods and systems for detecting protein interactions and detecting protein structures

The system addresses the challenge of varying cleavage efficiency in chemical cross-linking mass spectrometry by using a precursor mass optimization module and scoring functions, enhancing the sensitivity and accuracy of peptide and protein structure identification.

CN119091968APending Publication Date: 2024-12-06THE HONG KONG UNIV OF SCI & TECH
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202410719377.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2023-06-06
Filing Date
2024-06-05
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

Existing chemical cross-linking mass spectrometry methods face challenges with varying cleavage efficiency of cross-linked peptides during collision-induced dissociation, leading to incomplete or no cleavage, which complicates the interpretation of tandem mass spectra and reduces sensitivity.

Method used

A system and method that includes a precursor mass optimization module, tandem mass spectrum scoring, and a protein score database to enhance the identification of cleavable and non-cleavable cross-linked peptides by using scoring functions and a feedback mechanism to improve peptide and protein structure analysis.

Benefits of technology

The system significantly enhances the sensitivity and accuracy of identifying protein-protein interactions and protein structures by improving the interpretation of tandem mass spectra, achieving up to three times the sensitivity of existing methods.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119091968A_ABST
    Figure CN119091968A_ABST
Patent Text Reader

Abstract

The invention provides a comprehensive system and method for detecting protein-protein interaction and exploring a protein structure. The system comprises a plurality of components, including a container for receiving cross-linked precursor peptides, a mass spectrometer for generating a single mass spectrum and a tandem mass spectrum, and modules for precursor quality optimization, tandem mass spectrum scoring, protein scoring database construction, feedback mechanism processing and other functions. By integrating these components, the system obtains related cross-linking data sets from proteins cross-linked with cleavable or non-cleavable cross-linking agents, further identifying the interactions between peptides and proteins and the protein structure.
Need to check novelty before this filing date? Find Prior Art

Description

CROSS-REFERENCE TO RELATED APPLICATIONS:

[0001] This application claims priority to U.S. Provisional Patent Application Serial No. 63 / 506,369, filed on June 6, 2023, the entire disclosure of which is incorporated herein by reference. Technical Field

[0002] The present invention relates to proteomics. More specifically, the present invention relates to cross-linking mass spectrometry. Background Art

[0003] Mass spectrometry (MS) has become an important tool for rapid identification of protein primary structure, which has greatly changed the study of biological and pathological phenomena. Among them, chemical cross-linking mass spectrometry (XL-MS) is an important application that helps to resolve the three-dimensional structure of proteins and reveal protein-protein interactions (PPIs). Given the role of protein structure and protein-protein interactions in complex cellular mechanisms, chemical cross-linking mass spectrometry has become an indispensable tool in this field.

[0004] In chemical cross-linking mass spectrometry experiments, one of the important components is the cross-linker, a chemical reagent that is added to the protein solution. Cross-linkers usually have two functional groups that can connect specific side chains of amino acids to form covalent bonds and establish spatial connections within proteins or between proteins that interact with each other. Cross-linkers are divided into two categories: non-cleavable or cleavable. In tandem mass spectrometry (MS2), non-cleavable cross-linkers will remain intact during the dissociation process, while cleavable cross-linkers will break during the dissociation process due to their breakable bonds, and produce reporter ions in the mass spectrum, reducing the complexity of the calculation. In essence, both types of cross-linkers can provide important information about protein structure and protein-protein interactions. The main difference is that cleavable cross-linkers will produce additional characteristic reporter ions.

[0005] However, both cleavable and non-cleavable chemical cross-linking mass spectrometry experiments encounter a major challenge: the cleavage efficiency of cross-linked peptides varies during collision-induced dissociation (CID). The varying degree of cleavage of cross-linked peptides often results in incomplete or no cleavage of the cleaved peptides, making many tandem mass spectra difficult to interpret or reducing their sensitivity.

[0006] There are many strategies to solve the problem of uneven fragmentation, including the use of electron transfer dissociation (ETD) tandem mass spectrometry, step high-energy collision dissociation (HCD) tandem mass spectrometry or collision-induced dissociation tertiary (MS3) mass spectrometry. These methods mainly improve the above problems through hardware performance, but do not actively improve their applications, so the sensitivity that can be achieved is usually only between 20% and 50%.

[0007] Therefore, the present invention addresses this need and provides a highly sensitive and accurate method for distinguishing cleavable and non-cleavable cross-linked peptides. Summary of the invention

[0008] The purpose of the present invention is to provide a system or method for solving the above technical problems.

[0009] According to the first aspect of the present invention, a system for detecting protein-protein interactions and probing protein structures is provided. More specifically, the system includes the following components: a container for receiving cross-linked precursor peptides produced after digestion of proteins cross-linked with a cleavable or non-cleavable cross-linking agent; a mass spectrometer for generating precursor ions and generating a single mass spectrum (MS1) to determine the mass spectrum of the ions by the charge state and mass of the ions, and further fragmenting at least one precursor ion to generate a tandem mass spectrum (MS2) to be analyzed; a precursor mass optimization module for analyzing the elution spectrum generated by the mass spectrometer, and extracting from the elution spectrum single mass spectra obtained before and after the tandem mass spectrum to be analyzed, and merging them to form a complete isotope cluster of the precursor ion; a tandem mass spectrum scoring module, which can match the tandem mass spectrum to be analyzed with the complete isotope cluster through a scoring function S(x) to detect and store the highest p candidate peptides in the tandem mass spectrum and their scores, wherein the tandem mass spectrum scoring module further The method comprises the following steps: retrieving the highest-scoring α-peptide and β-peptide in the tandem mass spectrum as valid peptides, and generating a score histogram; a protein score database, which is constructed based on the score histogram and the valid peptides, and is used to provide overall information that can connect peptides to proteins, wherein the protein score database includes protein sequences in a biological system; a feedback mechanism processor, which is used to retrieve the highest p candidate peptides stored in the protein score database, and identify at least one candidate peptide in the protein score database, wherein the candidate peptide with the largest protein score in the at least one candidate peptide is defined as the matching result of the tandem mass spectrum to be analyzed, and the interaction between peptides or the protein structure can be identified from the matching result; and a cross-linking data set, which includes cleavable cross-linking agent data or non-cleavable cross-linking agent data, and the two types of data are obtained by cross-linking proteins in a biological system with cleavable or non-cleavable cross-linking agents, respectively.

[0010] According to an embodiment of the present invention, the system may further include a local alignment module, which may provide a series of compensation matches when there are no available characteristic ions in the tandem mass spectrum to be analyzed.

[0011] According to one embodiment of the present invention, the complete isotope cluster selected and calibrated by the precursor mass optimization module has a theoretical isotope cluster, and the single isotope mass is obtained by a Pearson correlation coefficient, and the correlation coefficient has the following formula: where x and y represent the theoretical isotope cluster and the complete isotope cluster, respectively.

[0012] According to an embodiment of the present invention, the following calculation formula is used for scoring: where k is the number of attachment sites in a single protein, is the protein vector, and the calculation results are summed to obtain the protein score (PS).

[0013] According to one embodiment of the present invention, the proteins cross-linked with a cleavable or non-cleavable cross-linking agent are further filtered, enriched and digested to generate cross-linked peptides.

[0014] According to an embodiment of the present invention, the biological system includes cells, tissues, blood, serum, and sputum.

[0015] According to an embodiment of the present invention, there is a preset correlation cutoff value set to 0.9 to accurately calculate the complete isotope cluster.

[0016] According to an embodiment of the present invention, the feedback mechanism processor uses a scoring function XlinkX as a scoring function for cleavable cross-link data: in ∈ is the tandem mass spectrum tolerance (ppm); when calculating the score of β peptides, f is equal to where α and β represent the α and β peptide masses, and f ABAC4 is the total number of fragment ions in the tandem mass spectrum.

[0017] According to an embodiment of the present invention, the feedback mechanism processor uses a scoring function Xcorr as a scoring function for non-cleavable cross-linking data: in is the vectorized form of the input tandem mass spectrum, and is the vectorized theoretical peptide sequence generated from the N-terminal ion (b-ion) and C-terminal ion (y-ion) of the fragmented precursor ion.

[0018] According to a second aspect of the present invention, there is provided a method for detecting protein-protein interactions and probing protein structures using the above system. More specifically, the method comprises the following steps: The cross-linked precursor peptides generated after digestion of proteins cross-linked with a cleavable or non-cleavable cross-linking agent are used as samples; performing mass spectrometry analysis of the sample to generate a plurality of precursor ions, and obtaining a single mass spectrum to determine the charge state and mass of the precursor ions; fragmenting at least one precursor ion among the plurality of precursor ions to obtain a tandem mass spectrum to be analyzed; Extracting single mass spectra obtained before and after the tandem mass spectrum to be analyzed from the elution spectrum, and forming a complete isotope cluster of the precursor ion through the precursor mass optimization module; Inputting the tandem mass spectrum into a tandem mass spectrum scoring module to match the tandem mass spectrum to be analyzed with the complete isotope cluster; Detect and store the top p candidate peptides and their scores of the α and β peptides of the tandem mass spectrum to be analyzed, where The tandem mass spectrum scoring module is used to further retrieve the highest-scoring α-peptide and β-peptide in the tandem mass spectrum to be analyzed, and the α-peptide and β-peptide are used as valid peptides to generate a score histogram; constructing the protein score database using the effective peptides and the score histogram; Retrieving the highest p candidate peptides of the stored α-peptide and β-peptide hits in the protein score database to identify at least one candidate peptide from the protein score database, wherein the at least one candidate peptide can be displayed in sequence form, and the candidate peptide with the largest protein score among the at least one candidate peptide is selected as a matching result of the input tandem mass spectrum; and The protein score is used to re-match the sequence to output cross-linked peptide mass spectrum matching results to further identify peptide interactions or protein structures.

[0019] According to one embodiment of the present invention, if the tandem mass spectrum matches the cleavable cross-linker data, the tandem mass spectrum is decharged.

[0020] According to an embodiment of the present invention, the method further includes a target decoy database to control output quality.

[0021] According to one embodiment of the present invention, the target decoy database is constructed by reversing protein sequences.

[0022] According to one embodiment of the present invention, the sample is further filtered, enriched and digested to generate cross-linked peptides.

[0023] According to one embodiment of the present invention, the search step further includes identifying the top 20 candidate peptides to facilitate subsequent adjustments after the matching is completed. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Embodiments of the present invention will be described in more detail below with reference to the accompanying drawings, in which:

[0025] Figures 1A-1D The operation flow of the precursor quality optimization module according to one embodiment of the present invention is shown, wherein Figure 1A Shows the generation of complete isotope clusters by merging upstream and downstream single mass spectra; Figure 1B Comparison of the isotope clusters of the merged single-shot mass spectra with the isotope clusters of the individual (original) single-shot mass spectra is shown, showing that the merged mass spectrum has a better structure and is more similar to the theoretical isotope clusters; Figure 1CDisplays a comparison of the theoretical isotope cluster to each possible position in a single mass spectrum and measures their similarity using the Pearson coefficient; and Figure 1D shows a system according to an embodiment of the present invention;

[0026] Figure 2A-2B The method flow of detecting protein-protein interaction and probing protein structure according to one embodiment of the present invention is shown, wherein Figure 2A Display the operation flow of the method; Figure 2B shows a protein feedback mechanism according to one embodiment;

[0027] Figure 3A-3B The performance of a system according to an embodiment of the present invention is shown. Figure 3A Shows comparison of mass spectrometry data of non-cleavable chemical cross-links between the system of the present invention (ECL 3.0) and existing analytical methods; Figure 3B Showing comparison of the system of the present invention and existing analytical methods for cleavable chemical cross-linking mass spectrometry data;

[0028] Figures 4A-4B Shows the comparison of the results of different systems using synthetic datasets. True positive results are represented by dark gray bars and false positives are represented by light gray bars. Figure 4A Displays the cross-linking pattern matching (CSM) and specific cross-linking results of the DSBU (disuccinimidyl dibutyric urea or BuUrBU) dataset; Figure 4B Displays cross-linking pattern matching and specific cross-linking results using the DSSO (disuccinimidyl sulfoxide) dataset;

[0029] Figure 5A-5B Shows a comparison of sensitivity and accuracy between different systems using a simulated data set, where Figure 5A Results of the sensitivity assessment are shown; Figure 5B The accuracy assessment results are shown; and

[0030] Figures 6A-6C The protein structure is verified using a system in one embodiment of the present invention, wherein Fig. 6A The mapping of the identified cross-linked peptides onto the X-ray diffraction structure of the GroEL complex (PDB ID: 1KP8) is shown; Figure 6B The display will Fig. 6A The structural model in the image is rotated 90 degrees; Figure 6C Displays the Cα-Cα distance histogram of all cross-linked peptides in the GroEL complex, with a distance cutoff of DETAILED DESCRIPTION

[0031] In the following description, systems and / or methods for detecting protein-protein interactions and probing protein structures, etc. are listed as preferred embodiments. It is obvious to those skilled in the art that modifications including additions and / or substitutions may be made without departing from the scope and spirit of the invention. In the following, specific details may be omitted to avoid obscuring the invention; however, this article is written to enable those skilled in the art to practice the technical content disclosed herein without excessive experimentation.

[0032] In order to solve the problem that it is difficult to accurately identify peptides with poor fragmentation in chemical cross-linking mass spectrometry (XL-MS), the present invention provides a system and / or method for peptide identification using additional information. In general, the introduction of a scoring function helps to find the correct peptide, but it cannot completely solve this problem. Therefore, the system / method provided by the present invention overcomes this limitation by combining additional information. Traditionally, database searches must match spectra through peptide sequence data in FASTA files, while ignoring the correlation between protein and peptide sequences, resulting in insufficient utilization of information in protein databases. For example, when two similar peptides have the same or very similar scores, but one of the peptides has many "brother" peptides from other tandem mass spectra, while the other lacks this type of identification, the former should be given priority as the correct matching result. Based on this concept, the present invention introduces a protein feedback process in the relevant identification system and method, and the sensitivity of the identification results is greatly improved compared with the prior art.

[0033] As used herein, the term "tandem mass spectrometry" refers to mass spectrometry analysis having two stages.

[0034] In chemical cross-linking mass spectrometry, it is usually necessary to obtain a single mass spectrum (MS1) and a tandem mass spectrum (MS2) spectrum. In the first step, the mass spectrometer scans all ions within a specific mass range (usually the full mass range of interest) and records their abundance (intensity) at each mass-to-charge ratio (m / z), thereby generating a single mass spectrum, which provides information about the ions present in the sample and their relative intensities. Therefore, a single mass spectrum helps to identify the precursor ions of interest for subsequent analysis.

[0035] Once the precursor ions of interest are identified from the single mass spectra, they are fragmented using collision-induced dissociation (CID) or other fragmentation methods to produce a spectrum of fragmented ions, i.e., a tandem mass spectrum. Tandem mass spectra provide information about the structure and composition of the precursor ions, and in the case of peptide analysis, amino acid sequence information. In chemical cross-linking mass spectrometry, tandem mass spectra help identify cross-linked peptides and examine their structures.

[0036] As used herein, the term "single-shot mass spectrum", also known as precursor ion mass spectrum, is the first stage of mass spectrometry analysis, which records the ion intensity corresponding to the intact precursor molecule (ion) in the sample, providing information about its abundance at the mass-to-charge ratio (m / z). Single-shot mass spectra help identify and quantify molecules in a sample by measuring their mass and abundance before fragmentation.

[0037] As used herein, the term "tandem mass spectrometry", also known as product ion spectrum, is a second-stage spectrum in mass spectrometry analysis, which records the results of fragment ions after the precursor ions are fragmented by collision-induced dissociation or other fragmentation methods, providing information about the molecular structure composition within the sample. Analysis of tandem mass spectrometry helps to understand the primary structure of the molecule and identify specific molecular fragments, so as to facilitate the identification and structural characterization of the compound.

[0038] As used herein, the term "elution spectrum" refers to the distribution of ions detected by a mass spectrometer as a function of elution time. The spectrum provides information about the retention times of different peptides, which is very important for the identification and peptide quantification steps in subsequent mass spectrometry analysis. In tandem mass spectrometry, the elution spectrum of a tandem mass spectrum refers to a collection of single mass spectra obtained before and after a specific tandem mass spectrum to be analyzed. By merging these single mass spectra, a complete isotope cluster can be formed for determining the mass of the precursor, which is beneficial for the subsequent steps of identifying the peptide. Therefore, the elution spectrum plays an important role in the optimization of precursor mass and is a key step in tandem mass spectrometry data analysis.

[0039] According to a first aspect of the present invention, a system for detecting protein-protein interactions and probing protein structure is provided.

[0040] refer to Figure 1D The system 10 includes multiple integrated components for comprehensive analysis. It includes a container 101 for receiving cross-linked precursor peptides generated after digestion of proteins cross-linked with cleavable or non-cleavable cross-linking agents, which can serve as the basis for subsequent analysis. A mass spectrometer 102 is used to generate precursor ions and generate a single mass spectrum, providing key information about the mass spectrum of the precursor ions and their charge state and mass. In addition, the mass spectrometer 102 will further fragment at least one precursor ion to generate a tandem mass spectrum of interest to be analyzed for subsequent analysis.

[0041] The precursor mass optimization module 103 analyzes the elution spectrum generated by the mass spectrometer 102, and extracts the single mass spectra obtained before and after the tandem mass spectrum to be analyzed from the elution spectrum, and then merges them to form a complete isotope cluster of the precursor ion, which is beneficial to improve the accuracy of subsequent analysis. The tandem mass spectrum scoring module 104 can use the scoring function S(x) to match the tandem mass spectrum to be analyzed with the complete isotope cluster, and detect and store the highest p candidate peptides and their scores in the tandem mass spectrum, and further retrieve the highest α peptide and β peptide hit points in the tandem mass spectrum as valid peptides, and generate a score histogram, which provides valuable information for peptide interactions.

[0042] In addition, the system 10 includes a protein score database 105 constructed based on the score histogram and valid peptides. The database provides overall information that can connect peptides to proteins, including protein sequences in biological systems. The feedback mechanism processor 106 retrieves the top p candidate peptides stored in the protein score database and identifies at least one candidate peptide, among which the candidate peptide with the largest protein score will be selected as the matching result of the tandem mass spectrum to be analyzed. This process can identify the interaction of peptides or protein structure from the matching results.

[0043] In addition, the system 10 comprises a cross-linking data set 107, including cleavable cross-linking agent data or non-cleavable cross-linking agent data, which are obtained by cross-linking proteins in biological systems with cleavable or non-cleavable cross-linking agents, respectively, and provide important information for comprehensive analysis.

[0044] In some embodiments, the system 10 also includes a specific module for optimizing and scoring the precursor mass to ensure the accuracy and efficiency of the analysis process. In addition, the system 10 has a preset correlation cutoff value and a scoring function to further enhance the system performance and enable a more robust analysis of protein-protein interactions and protein structure. For example, the system can include a local alignment module (not shown) that can provide a series of compensatory matches when there are no characteristic ions available on the tandem mass spectra to be analyzed.

[0045] The precursor mass optimization module 103 selects and calibrates a complete isotope cluster with a theoretical isotope cluster to obtain a single isotope mass through a Pearson correlation coefficient, and a specific calculation formula (described later) is used in the scoring process. The protein cross-linked with a cleavable or non-cleavable cross-linking agent is further filtered, enriched and digested to produce a cross-linked peptide; in addition, the above-mentioned biological system may include cells, tissues, blood, serum or sputum.

[0046] The feedback mechanism processor 106 uses the scoring function XlinkX as the scoring function for the cleavable cross-link data, and on the other hand, uses the scoring function Xcorr as the scoring function for the non-cleavable cross-link data.

[0047] According to a second aspect of the present invention, a method for detecting protein-protein interactions and probing protein structures using the above system is provided. The method comprises multiple steps for comprehensive analysis.

[0048] refer to Figure 2A First, a cross-linked precursor peptide produced after digestion of a protein cross-linked with a cleavable or non-cleavable cross-linking agent is used as a sample, and mass spectrometry is performed on it to generate precursor ions, and the obtained single mass spectrum can be used to determine the charge state and mass of the precursor ions. Subsequently, at least one precursor ion is fragmented, and the tandem mass spectrum to be analyzed is used for subsequent analysis.

[0049] The method also includes extracting single mass spectra obtained before and after the tandem mass spectrum to be analyzed from the elution spectrum, and forming a complete isotope cluster of the precursor ion through the precursor mass optimization module. This step can form a complete isotope cluster of the precursor ion, improving the accuracy of subsequent analysis. The tandem mass spectrum to be analyzed is input into the tandem mass spectrum scoring module to match it with the complete isotope cluster. Next, the highest p candidate peptides and their scores of the α and β peptides in the tandem mass spectrum to be analyzed are detected and stored, wherein the highest-scoring α peptide and β peptide in the tandem mass spectrum to be analyzed are further retrieved through the tandem mass spectrum scoring module, and used as valid peptides to generate a score histogram.

[0050] In addition, the method also includes constructing a protein score database by using valid peptides and score histograms to provide important information for subsequent analysis. For example, the highest p candidate peptides of stored α peptides and β peptides are retrieved from the protein score database to identify at least one candidate peptide from the protein score database, wherein the at least one candidate peptide can be displayed in sequence form, and the candidate peptide with the largest protein score among the candidate peptides is selected as the matching result of the input tandem mass spectrum. Subsequently, the sequence is re-matched using the protein score to output the cross-linked peptide mass spectrum matching result to further identify peptide interactions or protein structures.

[0051] In a specific embodiment, if the tandem mass spectrum matches the cleavable cross-linker data, it is decharged and a target decoy database is used to control false positive results, wherein the target decoy database is constructed by reversing the protein sequence. In addition, the sample can be further filtered, enriched and digested to produce cross-linked peptides; and in the step of searching, the method can further include identifying the top 20 candidate peptides to facilitate subsequent adjustment after the match is completed.

[0052] Precursor mass optimization is a key stage in the tandem mass spectrometry system of the present invention, which ensures accurate precursor mass information and helps maintain the accuracy of subsequent peptide identification. In this process, the system will roughly estimate the theoretical isotope distribution and compare it with the obtained single mass spectrum to determine the correctness of the obtained precursor mass. Figure 1A As shown in FIG. 1 , the workflow of this module includes extracting and merging single mass spectra near the current tandem mass spectrum to generate a merged spectrum, for example, extracting four single mass spectra located upstream and three single mass spectra located downstream. This step can also limit the scanned range. Figure 1B As shown in Figure 2, by comparing the fused isotope cluster with the theoretical isotope cluster, it can be found that the two have a higher similarity, showing a significant improvement compared to before the fusion. Figure 1C As shown, the system evaluates each possible position of the theoretical isotope cluster in the new single mass spectrum and calculates the Pearson correlation coefficient using Formula 1. The obtained coefficient represents the most suitable precursor mass. If the obtained coefficient does not exceed the preset threshold (usually 0.9), the system will automatically use the precursor weight provided in the RAW file. where x and y represent the theoretical isotope cluster and the complete isotope cluster, respectively.

[0053] In the next step, the scoring function S(x) is used to pair the mass spectra, and the top p candidate peptides and their scores in the tandem mass spectra are detected and stored for further evaluation. In a specific embodiment, three main scoring functions are provided as scoring functions for cleavable cross-linking data: the scoring function XlinkX in Formula 2, the normalized scoring function MeroX in Formula 3, and the scoring function Xcorr(SEQUEST) in Formula 4; for non-cleavable cross-linking data, the scoring function Xcorr(SEQUEST) is used. in ∈ is the tandem mass spectrum tolerance (ppm); when calculating the beta peptide fraction, f is equal to where α and β represent the α and β peptide masses, and f ABAC4 is the total number of fragment ions in the tandem mass spectrum. where k is the number of known ions with high abundance (≥10% intensity); i is the number of peaks with high abundance in the tandem mass spectrum; h is the number of known ions; d is the total number of fragment peaks in the tandem mass spectrum; and F int It is the ratio of the intensity of known ions to the intensity of total ions in a tandem mass spectrum. in is the vectorized form of the input tandem mass spectrum, and is the vectorized theoretical peptide sequence generated from the N-terminal ion (b-ion) and the C-terminal ion (y-ion) of the fragmented precursor ion.

[0054] After scoring, the algorithm continues to retrieve the highest-scoring peptides for each spectrum and generates a score histogram. The scoring function must be able to reflect the characteristics that high scores equal peptide matches with high accuracy. To achieve this, an indicator function is used to classify peptides in a binary manner by their scores, where the significant area is twice the median or 98th percentile of the score. Subsequently, all valid peptides are used to construct a protein scoring database.

[0055] The concept of protein scoring database is introduced as follows: through bioinformatics calculations, the proteins in the FASTA file are theoretically split into peptides; then these peptides containing connection sites are used to construct vectors corresponding to single proteins, and the length of the vector is equivalent to the number of peptides generated. After identifying the specific peptide of the protein m times in the effective area of ​​the histogram, the corresponding position in the vector is assigned a value of m. After calculating all the proteins in the FASTA file, the function C(x) in formula 5 is used to convert the counts and sum up all the count results from the protein vector to obtain the final protein score (expressed as PS). Unlike peptide databases used to identify potential peptide sequences in mass spectra, protein scoring databases provide comprehensive information to establish the association between peptides and proteins, mainly through theoretical protein sequences and scoring functions to establish the database. where k is the number of attachment sites in a single protein, is the protein vector, and the calculation results are summed to obtain the protein score (PS).

[0056] The subsequent steps involve a feedback process, which re-searches the stored top sites and cross-references them with candidate peptides in the protein scoring database. In the candidate peptide library, the peptide that exhibits the highest protein score is selected as the matching peptide for that mass spectrum.

[0057] Finally, Formula 6 is used to further introduce the target decoy false discovery rate (FDR) to adjust the false positive results. Where T represents the target hit point and D represents the bait hit point.

[0058] In general, the feedback mechanism is a set of conventional steps. First, the input tandem mass spectra are matched to the peptide sequences through the scoring function S(x), which stores the top p candidate peptides for each mass spectrum during the matching process; then, the first α and β peptide hits of all mass spectra are retrieved, and the histogram of the score distribution is plotted in the next step; the protein scoring database is built using the selected peptides (valid peptides) through the indicator function (where 2M d is twice the median); each protein in the database is digested into peptides, and only peptides containing the connection site are used to build the vector of its protein; next, the protein score is calculated using the counting function C(x); in the last step, the stored top p peptides are retrieved, and the one with the highest protein score is selected from these candidate peptides as the matching result and output.

[0059] refer to Figure 2B , protein T can be digested into t different peptide sequences, so its vector length is also t. In the histogram, sequence pep1 was identified twice, while sequence pep2 was only identified once in the significant region, and the vector of protein T was initialized accordingly (as shown in the protein score database). After counting all proteins in the FASTA file, the counting function C(x) is operated to transform the resulting counts, and all counting results from the protein vector are added to obtain the final protein score (PS). If there are 20 connection sites in the sequence of protein T, the final score can be calculated as 2.92. Unlike the peptide database used to provide potential peptide sequences for mass spectra identification, the protein score database established by theoretical protein sequences and scoring functions can provide overall information connecting both peptides and proteins. The last step is the feedback step, which is to return to the stored highest hit point and select candidate peptides in the protein score database, and use the candidate peptide with the largest protein score as the peptide matching the mass spectrum. By testing the whole proteome dataset, it is found that the performance of the scoring functions (SXlinkX, SMeroX and SXcorr) can be generally improved by 29.7% by introducing the protein feedback step.

[0060] Example

[0061] Example 1: Evaluation using synthetic datasets

[0062] Mass spectrometry cleavable data sets were downloaded from the synthetic cross-linked peptide library (PXD014337), and 95 synthetic peptides were cross-linked by DSSO and DSBU. These data sets were analyzed using PD 2.4MS Annika, PD 2.4XlinkX (60-day free version), MaxLynx (MaxQuant 2.0.3.0) and MeroX 2.0.1.4 and the system of the present invention (ECL-PF), wherein the local alignment module and the precursor mass optimization module were used in the ECL-PF system.

[0063] Figures 4A-4B The results of cross-link pattern matching (CSM) and the number of unique cross-links when the false discovery rate is 5% are shown in Figure 2. In the results of cross-link pattern matching, the ECL-PF system obtains the highest true positives. On average, the number of cross-link pattern matches identified by the ECL-PF system is 1.5 times that of the DSBU dataset ( Figure 4A ) and 3.0 times that of other software on the DSSO dataset ( Figure 4B ). For special cross-linking, due to the limited number of synthetic peptides in the sample, the ECL-PF system did not significantly outperform other software, but its performance was comparable to the best software. By observing the false discovery rate, it can be found that MS Annika has a high proportion of false positives on the DSSO dataset, and almost a quarter of the special cross-linking matching results are wrong; at the same time, the XlinkX software has the worst performance in all datasets. Therefore, MSAnnika and XlinkX will not be used in subsequent comparisons.

[0064] Example 2: Evaluation using a mixed simulated dataset

[0065] Due to the unique characteristics of chemical cross-linking mass spectrometry, a simulated data set was generated based on a real experimental data set for evaluation and testing. First, the signal peaks of α and β peptides were separated from the interference peaks in each identified spectrum, and the α peak of one spectrum was merged with the β peak of another spectrum to generate a new spectrum. Therefore, N spectra can generate N 2 A simulated spectrum with real data is generated. In addition, the isolated interfering peaks can be added to the new spectrum to further enhance its realism. These synthetic cross-linked peptide libraries are used to generate simulated data sets and their interfering peak ratios are adjusted to make the spectrum more complex.

[0066] Since MaxLynx can only input RAW files, it is not possible to run simulated data sets. Therefore, in this section, only the comparison between the ECL-PF system and MeroX is performed. For fairness, simulated data sets were generated using spectra that both software could recognize, and in order to test the robustness of both, the interference peaks extracted from the original spectra were also added to the simulated data. Finally, a total of 25,221 spectra of DSSO cross-linker and 84,741 spectra of DSBU cross-linker were generated, and the proportion of interference peaks ranged from 0% to 100%, where 100% means that all interference peaks from the original spectra were added to the new spectra. Figure 5A and Figure 5B The comparison of sensitivity and accuracy between the ECL-PF system and MeroX is shown. The results show that when there are no interfering peaks, the ECL-PF system recognizes almost all cross-linking spectra, while MeroX only recognizes half on average; as the proportion of interfering peaks increases, the sensitivity of both will decrease. In the worst case, the ECL-PF system can still recognize more than 80% of the spectra, but MeroX can only recognize 20%, but both can maintain a false discovery rate of less than 5%. This simulation result shows that compared with MeroX, the ECL-PF system can not only obtain better results, but also has better robustness to interfering peaks.

[0067] Example 3: Evaluation using real datasets

[0068] Four real data sets were used to evaluate the performance of the ECL-PF system, MaxLynx, and MeroX. The crosslinkers used could be selected from DSSO and DSBU to DSBSO. In the comparison, the false discovery rate of the four groups was set to 1%, and the precursor quality optimization module and the local alignment module were enabled in the ECL-PF system. The comparison results are shown in Table 1, which show that for all the tested data, the ECL-PF system greatly outperformed the other groups of data in terms of the number of crosslinking pattern matches and the portion of special crosslinks. It is worth noting that in the comparison of the real experimental data sets, the ECL-PF system identified 91.6% more crosslinking pattern matches and 52.6% more special crosslinks than other methods on average.

[0069] Table 1. Comparison of Maxlynx, MeroX and ECL-PF systems in processing real data sets

[0070] Example 4: Verification Structure

[0071] The results of the E. coli GroEL complex were extracted from the real experimental data set processed by the ECL-PF system, the X-ray diffraction structure of GroEL (PDB ID: 1KP8) was downloaded, and the cross-linking results calculated by the system of the present invention were mapped to the protein structure. A total of 101 pairs of lysine-lysine sites were identified from the ECL-PF system, of which 93.1% of the Cα-Cα distances met the restrictions of the cross-linker. The histogram of protein structure and distance is shown in Figures 6A-6C In particular, Fig. 6A The identified crosslinks (black lines) mapped on the X-ray diffraction structure of the GroEL complex (PDB ID: 1KP8) are shown, and a 90 degree rotation of the structural model is shown in Figure 6B shown; Figure 6C The Cα-Cα distance histogram of all crosslinks within the GroEL complex is shown, with a distance cutoff of (dashed line). Overall, this structural study shows the credibility of the ECL-PF system.

[0072] The present invention discloses that utilizing information about protein-peptide binding during identification can significantly improve identification sensitivity and also improve the primary scoring function in chemical cross-linking mass spectrometry. By implementing a protein feedback mechanism, the sensitivity of analysis of non-cleavable and cleavable cross-linking data can be increased by about three times (on a synthetic data set with a controllable false positive ratio).

[0073] Therefore, the system and method of the present invention have the opportunity to be applied to different fields such as drug development, biotechnology development, agriculture and food industry. For drug development, the system can be used as a platform for identifying new drug targets and screening and optimizing candidate drugs. In addition, since the system of the present invention is conducive to identifying protein-protein interactions, this is crucial for understanding disease mechanisms and developing effective treatments. In the field of biotechnology, the system can optimize the production of protein-related treatments (such as monoclonal antibodies) and help develop biosensors and other protein-based diagnostic tools. In addition, the system can also contribute to agricultural progress by studying protein-protein interactions in plants and developing disease-resistant crop varieties. In addition, the system can improve the quality and safety of food by identifying the structure and interaction of proteins in food.

[0074] In addition, the system can examine the three-dimensional structure of proteins by identifying intra- and intermolecular cross-linked peptide bonds, information that is critical for understanding protein folding, conformational changes, and interactions with other molecules.

[0075] Chemical cross-linking mass spectrometry is valuable for studying protein-protein interactions (PPIs) within biological systems. Through the identification of cross-linked peptides, the system can identify the proximity and spatial arrangement of proteins in complexes, providing insights into cell signaling, molecular recognition, and disease mechanisms.

[0076] In addition, chemical cross-linking mass spectrometry can be used to identify biomarkers in various diseases. For example, the system can identify specific cross-linked peptides associated with a disease by analyzing protein complexes and interactions in biological samples, which has the potential to provide diagnostic tools and therapeutic targets for subsequent development.

[0077] Chemical cross-linking mass spectrometry has a wide range of applications in biomedical research, including studies of protein dynamics, post-translational modifications, and cell signaling. The system of the present invention enables researchers to study complex biological processes at the molecular level, facilitating research in areas such as cancer biology, neurodegenerative diseases, and infectious diseases.

[0078] The above description has been presented to illustrate and describe the present invention, but is not intended to be exhaustive or to limit the present invention to the precise form disclosed. Many modifications and variations will be apparent to those skilled in the art.

[0079] The above embodiments were chosen and described in order to best explain the principles of the invention and its practical application, thereby enabling others skilled in the art to understand the invention for various embodiments and with various modifications as are suited to the particular use contemplated.

Claims

1. A system for detecting protein-protein interactions and probing protein structures, characterized in that: include: a container for receiving a cross-linked precursor peptide produced by digestion of a protein cross-linked with a cleavable or non-cleavable cross-linking agent; A mass spectrometer for generating at least one precursor ion and a single mass spectrum (MS1) to determine the mass spectrum of the ion by the charge state and mass of the ion, wherein the mass spectrometer can further fragment the at least one precursor ion to generate a tandem mass spectrum (MS2) to be analyzed; a precursor mass optimization module, for analyzing the elution spectrum generated by the mass spectrometer, extracting from the elution spectrum single mass spectra before and after the tandem mass spectrum to be analyzed, and merging them to form a complete isotope cluster of the at least one precursor ion; A tandem mass spectrum scoring module, which can calculate the scoring function S(x) to match the tandem mass spectrum to be analyzed with the complete isotope cluster to detect and store the highest p candidate peptides and their scores in the tandem mass spectrum, wherein the tandem mass spectrum scoring module further retrieves the highest-scoring α peptide and β peptide in the tandem mass spectrum as valid peptides and generates a score histogram; A protein score database constructed based on the score histogram and the effective peptides, and used to provide overall information that can connect peptides to proteins, wherein the protein score database includes protein sequences in a biological system; a feedback mechanism processor, configured to retrieve the top p candidate peptides stored in the protein score database, and identify at least one candidate peptide in the protein score database, wherein the candidate peptide having the maximum protein score among the at least one candidate peptide is the pairing result of the tandem mass spectrum to be analyzed, and the interaction between peptides or the protein structure can be identified from the pairing result; as well as The cross-linking data set includes cleavable cross-linking agent data or non-cleavable cross-linking agent data, which are obtained by cross-linking the proteins in the biological system with cleavable cross-linking agents or non-cleavable cross-linking agents, respectively.

2. The system according to claim 1, wherein the system further comprises a local alignment module that can provide a series of compensation pairs when there are no available characteristic ions in the tandem mass spectrum to be analyzed.

3. The system according to claim 1, wherein the complete isotope cluster selected and calibrated by the precursor mass optimization module has a theoretical isotope cluster, and the single isotope mass is obtained by a Pearson correlation coefficient, wherein the correlation coefficient has the following formula: Wherein x and y represent the theoretical isotope cluster and the complete isotope cluster, respectively.

4. The system according to claim 1, wherein the scoring is performed using the following formula: where k is the number of attachment sites in a single protein, is the protein vector, and the calculation results are summed to obtain the protein score (PS).

5. The system of claim 1, wherein the proteins cross-linked with cleavable or non-cleavable cross-linkers are further filtered, enriched and digested to produce cross-linked peptides.

6. The system of claim 1, wherein the biological system comprises cells, tissues, blood, serum, sputum.

7. The system of claim 2, wherein the precursor mass optimization module has a preset correlation cutoff value set to 0.9 to accurately calculate the complete isotope clusters.

8. The system according to claim 1, wherein the feedback mechanism processor uses the following scoring function XlinkX as a scoring function for cleavable crosslink data: in ∈ is the tandem mass spectrum tolerance (ppm); when calculating the score of β peptides, f is equal to where α and β represent the α and β peptide masses, and f ABAC4 is the total number of fragment ions in the tandem mass spectrum.

9. The system of claim 1, wherein the feedback mechanism processor uses the following scoring function Xcorr as a scoring function for the non-cleavable cross-link data: in is the vectorized form of the input tandem mass spectrum, and is the vectorized theoretical peptide sequence generated from the N-terminal ion (b-ion) and C-terminal ion (y-ion) of the fragmented precursor ion.

10. A method for detecting protein-protein interactions and probing protein structures using the system according to claim 1, characterized in that: include: After digestion of proteins cross-linked with cleavable or non-cleavable cross-linkers, the cross-linked precursor peptides obtained were used as samples; performing mass spectrometry analysis of the sample to generate a plurality of precursor ions, and obtaining a single mass spectrum to determine the charge states and masses of the plurality of precursor ions; fragmenting at least one precursor ion of the plurality of precursor ions to obtain a tandem mass spectrum to be analyzed; Extracting single mass spectra obtained before and after the tandem mass spectrum to be analyzed from the elution spectrum, and calculating the complete isotope cluster of the at least one precursor ion through the precursor mass optimization module; Inputting the tandem mass spectrum to be analyzed into the tandem mass spectrum scoring module to match the tandem mass spectrum to be analyzed with the complete isotope cluster; Detecting and storing the highest p candidate peptides and their scores of the α and β peptide hit points of the tandem mass spectrum to be analyzed, wherein the tandem mass spectrum scoring module further retrieves the highest-scoring α peptide and β peptide in the tandem mass spectrum to be analyzed, defines them as valid peptides, and generates a score histogram; Constructing the protein score database through the effective peptides and the score histogram; Retrieving the highest p candidate peptides of the α-peptide and the β-peptide stored in the protein score database to identify at least one candidate peptide from the protein score database, wherein the at least one candidate peptide can be displayed in a sequence form, and the candidate peptide with the maximum protein score among the at least one candidate peptide is selected as a matching result of the input tandem mass spectrum; as well as The sequences are re-matched by the protein scores, and the cross-linked peptide mass spectrum matching results are output to further identify peptide interactions or protein structures.

11. The method of claim 10, wherein if the tandem mass spectrum matches the cleavable cross-linker data, the tandem mass spectrum is decharged.

12. The method of claim 10, wherein the method further comprises a target decoy database to control false positive results.

13. The method according to claim 12, wherein the target decoy database is constructed by reversing protein sequences.

14. The method of claim 10, wherein the sample is further filtered, enriched and digested to produce cross-linked peptides.

15. The method according to claim 10, wherein in the searching step, it further comprises identifying the top 20 candidate peptides to facilitate subsequent adjustment after matching is completed.

Citation Information

Patent Citations

  • Semi-supervised learning method for evaluating randomness of biomolecular cross-linked mass spectrometry identification

    CN112151109A

  • Mass spectrum-based complex cross-linked peptide fragment identification method

    CN113571129A