Method, system and device for predicting pathogenicity of mutations based on sigma+

Abstract

The present application relates to a SIGMA+ prediction mutation pathogenicity analysis method, system and device. It comprises: obtaining a mutation of a gene to be predicted; mapping the mutation of the gene to be predicted onto a protein structure of the gene to be predicted to obtain a protein structure with the mutation mapped; inputting the protein structure with the mutation mapped into a SIGMA model to obtain a structure feature score of the gene to be predicted; inputting the mutation of the gene to be predicted into a predictor to obtain a variation effect score of the gene to be predicted; and inputting the structure feature score and the variation effect score of the gene to be predicted into a classifier to obtain a classification result of the mutation being a pathogenic mutation or a benign mutation. The method aims to combine SIGMA with other predictors based on the mutation of the gene to be predicted and the protein structure information with the mutation mapped, to explore the high prediction ability and potential application value of SIGMA for mutation pathogenicity.

Description

Technical Field

[0001] This invention relates to the field of gene data analysis, and more specifically, to an analytical method, apparatus, system, computer-readable storage medium, and application of SIGMA+-based prediction of mutation pathogenicity. Background Technology

[0002] The human genome contains approximately 3.16 billion base pairs, encoding about 30,000 genes, with an average of one variation occurring every 500-1000 base pairs. Within the human genome, protein-coding regions alone contain vast inter-individual differences; to date, 6.5 million missense variants have been observed. In fact, aside from the lethal protein locations in the genomes of the 8 billion people living on Earth, every protein location is potentially susceptible to variation. Most missense variants, which are clinically relevant, have highly variable effects on protein structure and function, but only a small fraction are pathogenic. To assess the pathogenicity of gene mutations, many state-of-the-art variant effect predictors (VEPs) have been developed, such as SIFT, REVEL, and EVE. However, assessing a large number of potential missense mutations (approximately 76 million in the human exome) using a single VEP is not optimal. Therefore, by integrating the advantages of multiple VEPs and obtaining information from multiple sources, including variant frequency, evolutionary conservation, and the physicochemical properties of amino acids, performance can be significantly improved.

[0003] Given the close relationship between protein structure and function, the structural context of mutations represents promising information independent of mutation frequency and evolutionary conservation. The scarcity of high-resolution 3D protein structures has hindered the development of variant evolution response (VEP) models utilizing protein structural information. However, the AlphaFold2 protein structure generation model expands the structural coverage of the human proteome from 17% to 98.5%, providing extensive and accurate structural information to aid in predicting variant effects. Therefore, integrating features extracted from multiple high-quality VEP models based on mutation and protein structure generation models for comprehensive pathogenicity assessment is of significant innovative value. Summary of the Invention

[0004] This invention integrates mutation information of the gene to be predicted and protein structure information of the predicted gene. Based on SIGMA, it explores potential applications in combination with other VEPs, deeply mines variation features in gene data, assesses the pathogenicity of mutations, and solves related life science problems. It has strong innovation in the life science field and will promote research on pathogenicity analysis of gene data.

[0005] This application discloses a pathogenicity analysis method based on SIGMA+ predicted mutations, including:

[0006] Obtain the mutation of the gene to be predicted, wherein the gene to be predicted includes wild-type genes;

[0007] The gene to be predicted is input into the protein structure generation model to obtain the protein structure of the gene to be predicted.

[0008] The mutation of the gene to be predicted is mapped onto the protein structure of the gene to be predicted, and the protein structure mapped with the mutation is obtained.

[0009] The protein structure mapped with mutations is input into the SIGMA model to obtain the structural feature score of the gene to be predicted.

[0010] The mutation of the gene to be predicted is input into the predictor to obtain the mutation effect score of the gene to be predicted.

[0011] The structural feature score and variation effect score of the gene to be tested are input into the classifier to obtain the classification result of whether the mutation is a pathogenic mutation or a benign mutation.

[0012] Furthermore, the protein structure generation model includes one or more of the following models: AlphaFold, AlphaFold2, and ProteoGAN; preferably, the protein structure generation model is AlphaFold2.

[0013] Furthermore, the SIGMA (Structure-Informed Germline Missense Mutation Assessor) model is developed based on the close relationship between protein structure and function, and is implemented using machine learning algorithms such as gradient boosting machines to evaluate the impact of missense variants in the context of protein structure. Specifically, the construction process of the SIGMA model includes feature selection, feature fusion, feature computation, and model optimization; the machine learning algorithms include any one or more of the following algorithms: GBDT, GBM, SVM, RF, Adaboost, and Apriori; the model optimization includes any one or more of the following methods: steepest descent method, Newton's method, and quasi-Newton method.

[0014] The feature selection process extracts features from the mapped mutated protein structure to obtain mutated protein structural features. These mutated protein structural features include thermodynamic features, protein volume features, secondary structure-related features, relative solvent accessibility features of mutated residues, and empirical rule features that can determine whether the mutation has a significant impact on the protein structure.

[0015] Furthermore, the predictor predicts the mutation effect of the gene to be predicted, and obtains the mutation effect score of the gene to be predicted; the predictor includes any one or more of the following software: Enformer, PROVEAN, EVE, DEOGEN2, MUTPRED, MutationAssessor, SIFT, ClinPred, PolyPhen2, REVEL; preferably, the predictor is PROVEAN, EVE, DEOGEN2, and MUTPRED;

[0016] Furthermore, the variation effect score of the gene to be predicted includes one or more of the following scores: sequence conservation score, allele mutation frequency score, missense mutation score, frameshift mutation score, nonsense mutation score, and insertion / deletion mutation score.

[0017] Preferably, the variation effect score of the gene to be predicted is a sequence conservation score, an allele mutation frequency score, and a missense mutation score.

[0018] Furthermore, the classifier employs one or more of the following algorithms: Boosting, XGBoost, and AdaBoost; preferably, the classifier employs Boosting.

[0019] A pathogenicity analysis device based on SIGMA+ predicted mutations, the device comprising: a memory and a processor;

[0020] The memory is used to store program instructions;

[0021] The processor is used to call program instructions, which, when executed, are used to perform the aforementioned pathogenicity analysis method based on SIGMA+ predicted mutations.

[0022] A pathogenicity analysis system based on SIGMA+ predicted mutations, comprising:

[0023] The acquisition module is used to acquire mutations in the gene to be predicted.

[0024] The structure generation module is used to map the mutation of the gene to be predicted onto the protein structure of the gene to be predicted, so as to obtain the protein structure mapped with the mutation.

[0025] The structural feature generation module is used to input the protein structure mapped with mutations into the SIGMA model to obtain the structural feature score of the gene to be predicted.

[0026] The mutation effect feature generation module is used to input the mutation of the gene to be predicted into the mutation effect predictor to obtain the mutation effect score of the gene to be predicted.

[0027] The classification module is used to input the structural feature score and variation effect score of the gene to be predicted into the classifier to obtain the classification result of whether the mutation is a pathogenic mutation or a benign mutation.

[0028] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described pathogenicity analysis method based on SIGMA+ predicted mutations.

[0029] Any of the following applications:

[0030] The above-mentioned equipment is used in the selection of auxiliary mutation pathogenicity analysis schemes; optionally, the treatment scheme includes mutation pathogenicity analysis and whether it is pathogenic due to changes in the type and number of gene mutation characteristics.

[0031] Application of the above-mentioned equipment in predicting the pathogenicity of mutations;

[0032] The application of the above-mentioned equipment in classifying mutant samples or predicting the properties of mutant samples;

[0033] The above system is used in the diagnosis of the occurrence and development of pathogenicity of mutations; optionally, the occurrence and development of pathogenicity of mutations are related to changes in the type and number of gene mutation characteristics;

[0034] Application of the above system in predicting the pathogenicity of mutations;

[0035] The above system is used in classifying mutant samples or predicting the attributes of mutant samples.

[0036] Advantages of this application:

[0037] 1. This application innovatively discloses a novel method for analyzing the pathogenicity of mutated genes. This method deeply mines the life laws hidden behind gene data, maps the mutation of the gene to be predicted to the wild-type protein structure, obtains the protein structure mapped with the mutation, and performs in-depth analysis of the protein structure characteristics based on this. At the same time, the mutation of the gene to be predicted is input into multiple high-quality VEPs, and the advantages of multiple VEPs are comprehensively considered to obtain the classification result of the mutation as a pathogenic mutation or a benign mutation, thereby improving the accuracy and depth of data analysis.

[0038] 2. This application innovatively integrates the variation effects of multiple features based on the structural features of the mutation of the gene to be predicted, including sequence conservation score, allele mutation frequency score, missense mutation score, frameshift mutation score, nonsense mutation score, and insertion / deletion mutation score.

[0039] 3. This application creatively discloses a pathogenicity analysis device and system based on SIGMA+ prediction of mutations. Through in-depth interpretation of gene mutation data, this application can be more accurately applied to fields such as auxiliary diagnosis of disease occurrence and development related to gene data, auxiliary selection of treatment plans, and prediction of mutation attributes. Attached Figure Description

[0040] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 This is a schematic flowchart of pathogenicity analysis based on SIGMA+ predicted mutations provided in an embodiment of the present invention;

[0042] Figure 2 This is a protein structure feature map based on AlphaFold2 that maps to mutations, provided in an embodiment of the present invention.

[0043] Figure 3 This is a prediction effect evaluation chart based on SIGMA pathogenicity analysis provided in an embodiment of the present invention;

[0044] Figure 4 This is the Spearman correlation analysis based on SIGMA-predicted mutations provided in the embodiments of the present invention;

[0045] Figure 5 This is a schematic diagram illustrating the importance features of mutation prediction based on SIGMA, provided in an embodiment of the present invention.

[0046] Figure 6 This is an interpretation of the variant effect score based on SIGMA prediction of mutations provided in the embodiments of the present invention;

[0047] Figure 7 This is the pathogenicity matrix for each amino acid substitution based on SIGMA prediction provided in this embodiment of the invention;

[0048] Figure 8 This is a pathogenicity effect analysis diagram based on SIGMA+ predicted mutations provided in an embodiment of the present invention;

[0049] Figure 9 This is a schematic diagram of a pathogenicity analysis device based on SIGMA+ predicted mutations provided in an embodiment of the present invention. Detailed Implementation

[0050] To enable those skilled in the art to better understand the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0051] In some of the processes described in the specification, claims, and accompanying drawings of this invention, multiple operations appearing in a specific order are included. However, it should be clearly understood that these operations may not be performed in the order they appear herein, or may be performed in parallel. The operation numbers, such as S101, S102, etc., are merely used to distinguish different operations and do not themselves represent any execution order. Furthermore, these processes may include more or fewer operations, and these operations may be performed sequentially or in parallel.

[0052] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] Figure 1 This is a schematic flowchart of a pathogenicity analysis method based on SIGMA+ prediction of mutations provided by an embodiment of the present invention. Specifically, the method includes the following steps:

[0054] S101: Obtain the mutation of the gene to be predicted.

[0055] In one embodiment, the gene to be predicted includes a wild-type gene, and obtaining mutations in the gene to be tested includes publicly available genetic data. Optionally, the mutation information of the gene to be tested also includes genetic information of the sample obtained by high-throughput sequencing. Optionally, publicly available genetic data includes one or more of the following datasets: dbSNP, dbVar, gnomAD, RefSeq, and ExAC, wherein dbSNP, dbVar, and RefSeq are all from the ClinVar database. A total of 27,165 benign and 22,957 pathogenic missense variants were retrieved from the gnomAD and ClinVar databases. For ClinVar mutations, single nucleotide missense mutations with an audit status of at least one star were retained (audited by an expert panel: provided criteria, multiple submitters, no conflict; provided criteria, single submitter).

[0056] Furthermore, obtaining the mutations of the gene to be predicted also includes preprocessing the mutation data. The preprocessing adopts one or more of the following methods: Optionally, using an algorithm to predict the effect of the mutation on splicing and excluding mutations with scores exceeding a threshold; Preferably, using the SpliceAI algorithm to predict the effect of the mutation on splicing and excluding mutations with SpliceAI scores greater than 0.5.

[0057] SpliceAI predicts splicing changes caused by single nucleotide variations, accurately predicting splicing sites (location and probability of abnormal splicing) from any mRNA precursor sequence, thus enabling the prediction of hidden splicing caused by mutations in non-coding RNA regions.

[0058] Optionally, the pretreatment also includes excluding mutations with more than a threshold number of residues; for example, proteins with more than 2,700 residues are excluded.

[0059] Optionally, pretreatment may include excluding mutations with clinical significance that is “contradictory to the explanation of pathogenicity”;

[0060] Optionally, preprocessing includes excluding mutations that cannot be mapped to the structure due to isomer inconsistencies.

[0061] In one specific embodiment, the mutated gene comprised 20,047 pathogenic (red) and 20,148 benign (blue) variants, and also included 27,928 mutations in six proteins characterized by the DMS experimental system, using 27 computational VEPs, such as EVE screening and DEOGEN2. The EVE score for each mutation was obtained from https: / / evemodel.org / . Predictions from other VEPs for each mutation were retrieved from the dbNSFP database (version 4.1a). These VEPs were divided into two categories: single predictors independent of other VEPs (n=16) and meta-predictors that integrated the results of other VEPs into input features (n=11).

[0062] S102: Map the mutation of the gene to be predicted to the protein structure of the gene to be predicted, and obtain the protein structure mapped with the mutation.

[0063] In one embodiment, the protein structure of the gene to be predicted is obtained by inputting the gene to be predicted into a protein structure generation model. Preferably, the gene to be predicted is a wild-type gene. The protein structure generation model refers to a model or software capable of generating the protein structure of a gene or polypeptide based on its amino acid sequence. Optionally, the protein structure generation model includes one or more of the following models: VAE, AlphaFold, AlphaFold2, and ProteoGAN; preferably, the protein structure generation model is AlphaFold2.

[0064] VAE (Variational Autoencoder) is a generative model that is based on joint probabilities.

[0065] AlphaFold accelerated and enabled large-scale discoveries, including deciphering the structure of the nuclear pore complex. AlphaFold predictions, with very high confidence, provided a nine-helix topology, offering clues about protein function that are crucial for unraveling the causes of rare genetic diseases.

[0066] AlphaFold2 discovered a novel protein-gated mechanism, glucose-6-phosphatase, an enzyme that identifies binding sites for inhibiting enzymes, and a transmembrane protein, Wolframin, located in the ER. It improved the accuracy of predictions to the atomic level, more quickly and accurately identifying enzyme active sites, with 350,000 predicted sites.

[0067] ProteoGAN, a conditional generative adversarial network, outperforms classic and recent deep learning baselines in protein sequence generation. It primarily expands protein screening with candidates that are farther from the known sequence space than previously possible, but is more likely to have functionality with relatively novel candidates than other methods.

[0068] In one specific embodiment, preferably, the protein structure of the gene to be predicted is a wild-type protein structure, obtained by inputting the wild-type gene into the AlphaFold2 protein structure generation model. The protein structure dataset in AlphaFold2 includes the AlphaFold protein structure database (AlphaFoldDB, https: / / alphafold.ebi.ac.uk / ), containing approximately 365,000 structure predictions. AlphaFold2 is used to provide these proteins with 1400 amino acid fragments; proteins with more than 2700 residues are excluded, as are mutations that cannot be mapped to a structure due to isoform inconsistencies.

[0069] S103: Input the protein structure mapped with mutations into the SIGMA model to obtain the structural feature score of the gene to be predicted.

[0070] In one embodiment, the SIGMA construction process is as follows: A mutation dataset containing labeling information is obtained; the mutation data is mapped onto protein structures to obtain mapped mutated protein structures; feature extraction is performed on the mapped mutated protein structures to obtain mutated protein structure features; machine learning algorithms are used to process the mutated protein structure features to obtain classification results; based on the predicted classification results and the actual results, the machine learning algorithm is optimized to obtain the predicted gene structure feature score, and the SIGMA model is output.

[0071] Optionally, the mutation dataset includes one or more of the following datasets: gnomAD, HumVar, ExoVar, PredictSNP, VariBench, SwissVar, Humsavar, and ClinVar;

[0072] Furthermore, obtaining the mutation dataset also includes preprocessing the mutation dataset, which may employ one or more of the following methods:

[0073] Optionally, an algorithm is used to predict the impact of mutations on splicing and exclude mutations with scores exceeding a threshold; preferably, the SpliceAI algorithm is used to predict the impact of mutations on splicing and exclude mutations with scores greater than 0.5.

[0074] Optionally, the pretreatment includes excluding mutations with a residue count exceeding a threshold; preferably, the threshold is 2700.

[0075] Optionally, preprocessing includes excluding mutations that cannot be mapped to the structure due to isomer inconsistencies.

[0076] Furthermore, the machine learning algorithms used in the SIGMA model primarily perform operations such as feature extraction, feature selection, feature fusion, feature computation, and model optimization. The machine learning algorithms used include one or more of the following: GBDT, GBM, SVM, RF, Adaboost, and Apriori algorithms; model optimization includes one or more of the following methods: steepest descent, Newton's method, and quasi-Newton methods.

[0077] In one embodiment, feature extraction is performed on the protein structure mapped with mutations to obtain 57 mutant protein structure features, including three levels of features: protein-level features, residue-level features, and mutation-level features. These 57 mutant protein structure features include thermodynamic features, protein volume features, secondary structure-related features, relative solvent accessibility features of mutant residues, and empirical rule features that can determine whether the mutation has a significant impact on the protein structure. Among these features, pathogenic mutations exhibit greater changes in protein stability than benign mutations.

[0078] Optionally, protein-level features include 16 thermodynamic features and 3 protein volume features; the thermodynamic features include the total unfolding energy of wild-type proteins and its 15 components such as van der Waals collision energy, hydrogen bond energy, and side chain entropy; the protein volume features include total volume, interstitial volume, and van der Waals volume.

[0079] Optionally, residue-level features include eight secondary structure-related features and the relative solvent accessibility (RSA) of the mutated residue, used to characterize the structural background of the mutated residue. For each mutated residue, its secondary structure is assigned using the DSSP (protein secondary structure dictionary) program, generating features corresponding to eight secondary structure types: DSSP(H), DSSP(E), DSSP(G), DSSP(I), DSSP(T), DSSP(S), and DSSP(C).

[0080] Optionally, mutation level features include 16 thermodynamic features describing changes in protein stability after mutation and 13 features derived from empirical rules for determining whether a mutation has a significant effect on protein structure; the 16 thermodynamic features include the free energy difference between wild-type and mutant proteins and their 15 components; the other 13 features are combinations of the structural background and physicochemical properties of the mutant / mutant residues, including glycine bending, proline in the α-helix, proline / glycine substitution of residues in the α-helix, cysteine ​​residues, energy charge loss, energy charge loss, energy hydrophobicity replacement, outlier substitution, disulfide bond breakage, buried salt bridge breakage, buried proline, energy displacement of glycine, unfolding free energy of wild-type protein, and unfolding free energy difference between wild-type and mutant proteins.

[0081] Figure 2 This study demonstrates the structural features of mutant proteins obtained by mapping mutations to wild-type protein structures generated based on AlphaFold2. The results show that these structural features are associated with pathogenicity, particularly relative solvent accessibility (RSA) and free energy difference (ΔΔG), which are significantly correlated with pathogenicity. Figure 2 As shown in Figure A, the RSA of pathogenic variants was significantly lower than that of benign variants (P<2.2e-16, Mann-Whitney U test), which is consistent with the fact that most proteins are less tolerant of hidden mutations than exposed mutations. Figure 2 As shown in B, the effect of ΔΔG on protein stability under single amino acid substitution was measured. Pathogenic variants showed greater changes in protein stability than benign variants (P < 2.2e-16, Mann-Whitney U test). Figure 2 As shown in Figure C, as expected, disulfide bond breaking has the highest positive predictive value among all structural features (odds ratio [OR] = 93.8, 95% confidence interval [CI] = 44.5–198, P < 2.2e-16, Pearson chi-square test). Almost all (98.72%) missense mutations that break disulfide bonds are pathogenic, supporting the important role of disulfide bonds in protein function. Figure 2As shown in D, the association between the type of secondary structure in which the mutation occurs and its pathogenicity is consistent with prior knowledge; mutations in the circular or irregular extensions of proteins (Dictionary of Protein Secondary Structures [DSSP]-C) are benign (OR = 0.32, 95% CI = 0.31–0.34, P < 2.2e-16, Pearson chi-square test). In contrast, mutations in regular secondary structures tend to be pathogenic, especially mutations in α-helices (DSSP-H; OR = 1.73, 95% CI = 1.66–1.79, P < 2.2e-16, Pearson chi-square test) or β-sheets (DSSP-E; OR = 1.97, 95% CI = 1.87–2.08, P < 2.2e-16, Pearson chi-square test).

[0082] In one specific embodiment, a final labeled dataset containing 27,165 benign (negative) and 22,957 pathogenic (positive) missense mutations is used as the "gold standard" dataset. The dataset is divided into 80% for training and 20% for testing. The gnomAD database retains 27,928 mutations for model evaluation. Training the SIGMA model can be as follows: obtain labeled samples of different categories from the dataset (e.g., pathogenic / potentially pathogenic variants are labeled positive, while benign / potentially benign variants are labeled negative); repeat the protein structure generation and feature extraction processes described above to obtain a feature set; input the feature set into the SIGMA model to obtain the classification result; calculate the loss between the classification result and the true value using a loss function; and then update the parameters using backpropagation and an optimizer to obtain the trained SIGMA model.

[0083] Furthermore, the labeling information was defined as follows: pathogenic / potentially pathogenic variants were labeled positive, while benign / potentially benign variants were labeled negative. For gnomAD mutations, preprocessing was performed using SpliceAI to predict the gnomAD database (for deep intron regions: distance from the exon-intron boundary >50 nt). Due to the difficulty of deep intron prediction, only 56% deletions were observed when the threshold was set to 0.8. Therefore, the impact of mutations on splicing was predicted, and mutations with SpliceAI scores greater than 0.5 were excluded. Mutations introducing mystery splicing sites primarily affected mRNA splicing rather than protein structure, and selected common missense mutations (the highest allele frequency in all populations with gnomAD > 0.05) were labeled negative. In addition, the GOF / LOF database labeled 193 gain-of-function (GoF) and 921 loss-of-function (LoF) pathogenic missense variants.

[0084] Furthermore, the SIGMA model was evaluated based on genomic data type to assess the association between the aforementioned structural features and variant pathogenicity. For dichotomous features, contingency tables were constructed, and the chi-square test was used to determine whether an association existed between the two dichotomous variables. For continuous features, logistic regression analysis was used to examine the association between the feature and variant pathogenicity. The strength of the association between the feature and variant pathogenicity was quantified using the odds ratio (OR) and 95% confidence interval (CI). All statistical analyses and data visualizations were performed using R software packages: h2o, caret, pROC, forestplot, ggpubr, ggsci, viridis, and cutpointr.

[0085] Specifically, such as Figure 3 The performance evaluation plot shown uses 40,195 mutations with structural features as the training dataset and employs a SIGMA model based on gradient boosting machine (GBM) to predict the pathogenicity of missense mutations. Figure 3 As shown in Figure A, the SIGMA score distribution for 20,047 pathogenic (red) and 20,148 benign (blue) variants in the training set is presented. The SIGMA scores, calculated using out-of-fold prediction, represent the quantitative pathogenicity of variants. The SIGMA scores (from 0 to 1) show a bimodal distribution, indicating that SIGMA has strong discriminative power. For the training set, Figure 3 The area under the ROC curve (AUC) of the receiver operating characteristic shown in B is 0.944 (95% CI = 0.942 - 0.946), indicating an incorrect prediction. Consistently, as... Figure 3 As shown in C, high prediction accuracy was achieved on the test set (AUC = 0.933, 95% CI = 0.928–0.938). Figure 3 D shows the AUC of SIGMA compared to 16 individual variation effect predictors (VEPs) using the test set. SIGMA outperformed all individual VEPs, with AUCs ranging from 0.779 (FATHMM) to 0.929 (MutPred). Figure 3 E represents the correlation between VEP and Deep Mutation Scan (DMS) measurements, where Spearman's correlation is calculated between the functional score of the DMS experiment, the predicted score of SIGMA, and 16 individual VEPs. The closest competitor is DEOGEN2 (overall rho = 0.387, Spearman correlation analysis); Figure 3 E) contains extensive heterogeneous information, and the performance ranking of the predictor EVE improved from sixth place on the labeled dataset to third place on the DMS dataset. Figure 3 E).

[0086] Furthermore, to limit data recurrence that could potentially overstate predictor performance, a deep mutation scanning DMS independent of the labeled dataset was used to assemble a test dataset for six proteins (BRCA1, P53, MSH2, PTEN, VKORC1, and HRAS). The study used high-throughput functional assays to quantify the functional impact of all possible mutations in the proteins / protein domains with continuous scoring, preserving a high-quality DMS study of 28,293 mutations across the six proteins. Figure 4 As shown, the correlation between SIGMA and DMS scores and high-quality DMS studies for six proteins is illustrated (Spearman correlation analysis), with the color of each hexagon representing the number of mutations. The results showed that SIGMA and the DMS dataset had the highest correlation among all individual predictors (overall rho = 0.419, Spearman correlation analysis), especially for BRCA1 (rho = 0.286), PTEN (rho = 0.519), and HRAS (rho = 0.404). Figure 3 E; Figure 4 ).

[0087] Figure 5 This demonstrates feature importance analysis based on the SIGMA model, from... Figure 5 A shows the ten most important features affecting the discriminative power of SIGMA. Among them, the residue-level feature RSA contributes the most to the discriminative power of SIGMA, followed by two mutation-level features (ΔΔG and ΔVander). Seven of these ten most important features are protein-level features, demonstrating the important role of wild-type protein stability in predicting the pathogenicity of variants. Figure 5 B and Figure 5 C represents the gene set enrichment analysis (GSEA) plots for ion channel genes (B) and collagen family genes (C), respectively. Notably, RSA alone can predict the pathogenicity of variants in 76 out of 200 genes with an accuracy >0.8 (with at least 10 pathogenic variants and 10 benign variants used for evaluation). Among these, RSA showed the highest classification accuracy for ion channel genes relative to solvent accessibility (P = 8.67e-04); however, RSA alone performed unsatisfactorily for 11% of the genes (21 out of 200 genes, accuracy <0.5), which are rich in the collagen gene family (P = 1.99e-06). Figure 5 C). Additionally, ΔΔG represents the difference in expansion free energy between the wild-type and mutant proteins; ΔG represents the expansion free energy of the wild-type protein; and VdW represents van der Waals.

[0088] S104: Input the mutation of the gene to be predicted into the predictor to obtain the variation effect score of the gene to be predicted.

[0089] In one embodiment, the predictor performs mutation effect prediction on the acquired mutations in the gene to be predicted, obtaining a mutation effect score for the gene to be predicted. The predictor includes any one or more of the following software: Enformer, PROVEAN, EVE, DEOGEN2, MUTPRED, MutationAssessor, SIFT, ClinPred, PolyPhen2, and REVEL; preferably, the predictor is PROVEAN, EVE, DEOGEN2, or MUTPRED.

[0090] PROVEAN is a tool for predicting whether protein sequence variations affect protein function. Based on evolutionary conservation, neural network models, and the BLOSUM62 amino acid substitution scoring matrix, it identifies whether non-synonymous mutations or InDels affect protein biological function. Prediction scores range from -14 to 14, with a threshold of -2.5. Scores of -14 to 2.5 are predicted as Deleterious, and scores of -2.5 to 14 are predicted as Neutral. Lower scores indicate more harmful mutations, and vice versa.

[0091] EVE estimates the likelihood of each single amino acid variation being benign or pathogenic, understands the pathogenicity of human missense variations from the distribution of sequence variations across species, outperforms other computational prediction models in predicting clinical outcomes, and scores as high as or better than the current gold standard high-throughput experiments for testing the impact of mutations on biological function.

[0092] DEOGEN2 contains heterogeneous information about the molecular effects of mutations, the domains involved, gene relevance, and the interactions they participate in. This extensive contextual information is non-linearly mapped to a single harmfulness score for each mutation.

[0093] MUTPRED's function is to predict the pathogenicity and molecular mechanisms of amino acid substitutions, using Fasta-formatted amino acid sequences as the primary input. MUTPRED is a collection of machine learning tools that can predict the pathogenicity of protein-coding mutations to infer the molecular mechanisms of diseases.

[0094] REVEL is an ensemble method for missense mutation prediction that combines prediction results from multiple software programs (MutPred, FATHMM, VEST, PolyPhen, SIFT, PROVEAN, MutationAssessor, MutationTaster, LRT, GERP, SiPhy, phyloP, and phastCons). It is trained using pathogenic and rare neutral nonsense mutations and shows good prediction performance for rare missense mutations. The prediction score for a single mutation ranges from 0 to 1, and the threshold can be set according to the required sensitivity and specificity.

[0095] Specifically, the variation effect score of the gene to be predicted includes one or more of the following scores: sequence conservation score, allele mutation frequency score, missense mutation score, frameshift mutation score, nonsense mutation score, and insertion / deletion mutation score; preferably, the variation effect score of the gene to be predicted includes sequence conservation score, allele mutation frequency score, and missense mutation score.

[0096] Figure 6 This represents the application of the predicted gene variant effect score in mutation interpretation SIGMA. The predicted gene variant effect score includes sequence conservation score, allele mutation frequency score, and missense mutation score. From Figure 6 It was found that mutations with high SIGMA scores tend to appear in highly conserved regions. Figure 6 A), with a low allele mutation frequency ( Figure 6 B), and exists in regions with poor tolerance to missense variants. Figure 6 C). Sequence conservation, allele mutation frequency, and missense score are accurate predictors of variant pathogenicity. Furthermore, SIGMA can effectively distinguish between pathogenic and benign variants within each subgroup. Figure 6 These results suggest that SIGMA may function in a complementary manner to enhance its ability to predict the pathogenicity of variants. Figure 6 The boxes in the figure show the median ± one quartile (1 / 4), with whiskers extended to 1.5 times the quartile range. The Mann-Whitney U test was used to evaluate SIGMA scores for benign and pathogenic variants; a two-sided P < 0.05 was considered statistically significant, and **** indicated P < 0.0001.

[0097] S105: Input the structural feature score and variation effect score of the gene to be predicted into the classifier to obtain the classification result of whether the mutation is a pathogenic mutation or a benign mutation.

[0098] In one embodiment, the structural feature score and variation effect score of the gene to be predicted are obtained through a combined predictor (SIGMA+), constructed using a resilient network penalized logistic regression and a five-fold cross-validation strategy based on the same training set. The classifier employs one or more of the following models: Boosting, XGBoost, and AdaBoost; preferably, the classifier employs Boosting.

[0099] Boosting is a mainstream and representative technique in ensemble learning in machine learning. It often beats deep neural networks' XGBoost in many data analysis competitions and is an efficient implementation of the GradientBoost algorithm in the Boosting family.

[0100] Figure 7 The pathogenicity matrix for each amino acid substitution is shown. According to... Figure 7 The calculated saturation mutation results for all disease-related genes shown are presented, using average SIGMA scores to measure the pathogenicity of each amino acid substitution. Colors range from yellow (low pathogenicity) to black (high pathogenicity). On average, variations are more likely to be pathogenic when wild-type residues are replaced by proline or when hydrophobic residues are replaced by hydrophilic residues. Pre-calculated SIGMA scores for all possible missense variants (n=48,635,743) in the disease-related genes (n=3,705) are published on an interactive online platform (https: / / www.sigma-pred.org / ).

[0101] The above method was used to assist in the selection of a gene mutation pathogenicity analysis scheme. Based on the gene information of the sample, the classification result of whether the mutation is pathogenic was predicted. Figure 8 This is a pathogenicity effect analysis diagram based on SIGMA+ predicted mutations provided in an embodiment of the present invention. Figure 8 A demonstrates the potential applications of combining SIGMA with other predictors. For example, combining SIGMA with four individual VEPs with high performance—DEOGEN2, EVE, PROVEAN, and MutPred—can yield more comprehensive predictors of variant pathogenicity. Various combinations of these five predictors enhance predictive performance (AUC range of 0.942–0.954). SIGMA+ is the overall combination of these five predictors, which significantly distinguishes between benign and pathogenic variants. Figure 8 B specifically demonstrated its high predictive ability for pathogenicity of variants (AUC = 0.966). Figure 8 C demonstrates the correlation between VEP and deep mutation scanning (DMS) measurements. From Figure 8 As can be seen, SIGMA contributes the most to the combination, while DEOGEN2 contributes the least, indicating that SIGMA has strong discriminative power. Therefore, mining additional information sources may be more beneficial than iteratively building meta-predictors using more and more existing predictors.

[0102] This invention provides a pathogenicity analysis system based on SIGMA+ predicted mutations, comprising:

[0103] The acquisition module is used to acquire mutations in the gene to be predicted.

[0104] The structure generation module is used to map the mutation of the gene to be predicted onto the protein structure of the gene to be predicted, so as to obtain the protein structure mapped with the mutation.

[0105] The structural feature generation module is used to input the protein structure mapped with mutations into the SIGMA model to obtain the structural feature score of the gene to be predicted.

[0106] The variant effect feature generation module is used to input the mutation of the gene to be predicted into the variant effect predictor to obtain the variant effect score of the gene to be predicted.

[0107] The classification module is used to input the structural feature score and variation effect score of the gene to be predicted into the classifier to obtain the classification result of whether the mutation is a pathogenic mutation or a benign mutation.

[0108] Figure 9 This invention provides a pathogenicity analysis device based on SIGMA-predicted mutations. The device includes a memory and a processor; it may also include an input device and an output device. The processor, memory, input device, and output device can be connected via a bus or other means. Figure 9 Taking the bus connection method as an example; the memory is used to store program instructions; the processor is used to call program instructions. When the program instructions are executed, they are used to perform the above-mentioned pathogenicity analysis method based on SIGMA+ prediction of mutations.

[0109] The present invention provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the above-described pathogenicity analysis method based on SIGMA+ prediction of mutations.

[0110] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0111] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, indirect coupling or communication connection between apparatuses or modules, and may be electrical, mechanical, or other forms.

[0112] The modules described as separate components may or may not be physically separate. Similarly, the components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0113] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0114] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, etc.

[0115] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware, and the program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk.

[0116] The computer device provided by the present invention has been described in detail above. For those skilled in the art, there will be changes in the specific implementation and application scope based on the ideas of the embodiments of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A pathogenicity analysis method based on SIGMA+ predicted mutations, comprising: Obtain the mutation of the gene to be predicted, wherein the gene to be predicted includes wild-type genes; The gene to be predicted is input into the protein structure generation model to obtain the protein structure of the gene to be predicted. The mutation of the gene to be predicted is mapped onto the protein structure of the gene to be predicted, and the protein structure mapped with the mutation is obtained. The protein structure mapped with mutations is input into the SIGMA model to obtain the structural feature score of the gene to be predicted. The mutation of the gene to be predicted is input into the predictor to obtain the mutation effect score of the gene to be predicted. The structural feature score and variation effect score of the gene to be predicted are input into the classifier to obtain the classification result of whether the mutation is a pathogenic mutation or a benign mutation; The SIGMA model is implemented based on machine learning algorithms, including feature selection, feature fusion, feature computation, and model optimization. The feature selection process extracts features from the mapped mutated protein structure to obtain mutated protein structural features. These mutated protein structural features include thermodynamic features, protein volume features, secondary structure-related features, relative solvent accessibility features of mutated residues, and empirical rule features that can determine whether the mutation has a significant impact on the protein structure. The construction process of SIGMA is as follows: Obtain a mutation dataset containing labeling information; map the mutation data onto protein structures to obtain the protein structures mapped with mutations; extract features from the protein structures mapped with mutations to obtain the mutant protein structure features; Machine learning algorithms are used to process the structural features of mutant proteins to obtain classification results. Based on the predicted classification results and the actual results, the machine learning algorithm is optimized to obtain the structural feature scores of the genes to be predicted, and the SIGMA model is output.

2. The pathogenicity analysis method based on SIGMA+ predicted mutations according to claim 1, characterized in that, The predictor performs mutation effect prediction on the gene to be predicted and obtains the mutation effect score of the gene to be predicted; the predictor includes any one or more of the following software: Enformer, PROVEAN, EVE, DEOGEN2, MUTPRED, MutationAssessor, SIFT, ClinPred, PolyPhen2, REVEL.

3. The pathogenicity analysis method based on SIGMA+ predicted mutations according to claim 2, characterized in that, The predictors are PROVEAN, EVE, DEOGEN2, and MUTPRED.

4. The pathogenicity analysis method based on SIGMA+ predicted mutations according to claim 1, characterized in that, The variation effect score of the gene to be predicted includes one or more of the following scores: sequence conservation score, allele mutation frequency score, missense mutation score, frameshift mutation score, nonsense mutation score, and insertion / deletion mutation score.

5. The pathogenicity analysis method based on SIGMA+ predicted mutations according to claim 4, characterized in that, The variation effect score of the gene to be predicted includes sequence conservation score, allele mutation frequency score, and missense mutation score.

6. The pathogenicity analysis method based on SIGMA+ predicted mutations according to claim 1, characterized in that, The machine learning algorithm includes one or more of the following algorithms: The algorithms include GBDT, GBM, SVM, RF, Adaboost, and Apriori; the model optimization includes any one or more of the following methods: steepest descent method, Newton method, and quasi-Newton method.

7. The pathogenicity analysis method based on SIGMA+ predicted mutations according to claim 1, characterized in that, The protein structure generation model includes one or more of the following generation models: AlphaFold, AlphaFold2, and ProteoGAN.

8. The pathogenicity analysis method based on SIGMA+ predicted mutations according to claim 1, characterized in that, The protein structure generation model is AlphaFold2.

9. The pathogenicity analysis method based on SIGMA+ predicted mutations according to claim 1, characterized in that, The classifier employs one or more of the following algorithms: Boosting, XGBoost, and AdaBoost.

10. The pathogenicity analysis method based on SIGMA+ predicted mutations according to claim 1, characterized in that, The classifier uses Boosting.

11. A pathogenicity analysis device based on SIGMA+ predicted mutations, characterized in that, The device includes: a memory and a processor; the memory is used to store program instructions; the processor is used to invoke the program instructions, and when the program instructions are executed, they are used to perform the pathogenicity analysis method based on SIGMA+ prediction of mutations as described in any one of claims 1-10.

12. A pathogenicity analysis system based on SIGMA+ predicted mutations, characterized in that, The system includes: an acquisition module for acquiring mutations in the gene to be predicted; The structure generation module is used to map the mutation of the gene to be predicted onto the protein structure of the gene to be predicted, so as to obtain the protein structure mapped with the mutation. The structural feature generation module is used to input the protein structure mapped with mutations into the SIGMA model to obtain the structural feature score of the gene to be predicted. The mutation effect feature generation module is used to input the mutation of the gene to be predicted into the mutation effect predictor to obtain the mutation effect score of the gene to be predicted. The classification module inputs the structural feature score and variation effect score of the gene to be predicted into the classifier to obtain the classification result of whether the mutation is a pathogenic mutation or a benign mutation; The SIGMA model is implemented based on machine learning algorithms, including feature selection, feature fusion, feature computation, and model optimization. The feature selection extracts features from the mapped mutated protein structure to obtain mutated protein structural features. The mutated protein structural features include thermodynamic features, protein volume features, secondary structure-related features, relative solvent accessibility features of mutated residues, and empirical rule features that can determine whether the mutation has a significant impact on the protein structure. The SIGMA construction process is as follows: A mutation dataset containing labeling information is obtained; the mutation data is mapped onto protein structures to obtain mapped mutated protein structures; feature extraction is performed on the mapped mutated protein structures to obtain mutated protein structure features; machine learning algorithms are used to process the mutated protein structure features to obtain classification results; based on the predicted classification results and the actual results, the machine learning algorithm is optimized to obtain the predicted gene structure feature scores, and the SIGMA model is output.

13. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the pathogenicity analysis method based on SIGMA+ prediction of mutations as described in any one of claims 1-10.

Images