DAMpred: Recognizing Disease-Associated nsSNPs through Bayes-Guided Neural-Network Model Built on Low-Resolution Structure Prediction of Proteins and Protein–Protein Interactions

Abstract

Nearly one-third of non-synonymous single-nucleotide polymorphism (nsSNPs) are deleterious to human health but recognition of the disease-associated mutations remains a significant unsolved problem. We proposed a new algorithm, DAMpred, to identify disease-causing nsSNPs through the coupling of evolutionary profiles with structure predictions of proteins and protein-protein interactions. The pipeline was trained by a novel Bayes-guided artificial neural network algorithm that incorporates posterior probabilities of distinct feature classifiers with the network training process. DAMpred was tested on a large-scale dataset involving 10,635 nsSNPs from 2,154 ORFs in the human genome and recognized disease-associated nsSNPs with an accuracy 0.80 and Matthew's correlation coefficient (MCC) 0.601 that is 9.1% higher than the best of other state-ofthe-art methods. In the blind test on the TP53 gene, DAMpred correctly recognized the mutations causative of Li-Fraumeni-like syndrome with an MCC that is 27% higher than the control methods. The study demonstrates an efficient avenue to quantitatively model the association of nsSNPs with human diseases from low-resolution protein structure prediction, which should find important usefulness in diagnosis and treatment of genetic diseases.

Circuit topology predicts pathogenicity of missense mutations

“…Early methods relied primarily on evolutionary conservation, where mutation of a well conserved residue is more likely to be pathogenic 4,52 . Recently developed, advanced methods also incorporate structural information and predictions 4,8,53 . In the case that it is nonredundant with other features, circuit topology information has the potential to improve state‐of‐the‐art predictors, as well as providing mechanistic insight.…”

“…Databases of pathogenic and benign mutations have been established, where the domain of the protein containing the mutation can often be linked to a crystal structure or NMR structure in the Protein Data Bank 1–5 . However, it is still a challenge to predict whether a given mutation will lead to disease, suggesting that common measures such as ∆∆ G , the change in free energy of folding upon mutation, 6 may not account for all the relevant information contained in the structure 7,8 . This “missing information” may include kinetic rates and misfolding propensity 9–13 .…”

Incorporating structural features to improve the prediction and understanding of pathogenic amino acid substitutions

“…Many studies have adopted sequence features and the predicted structural features based on sequence, but only a few directly extract features from protein 3D structures, whether experimental structures or homology models [45,57,59,60]. According to the paradigm of sequencestructure-function, protein structures are more directly related to function than sequence, so the 3D structures should be able to provide greater understanding of pathogenic AASs.…”

“…Similar results were obtained on TestDataset 1 (0.244 vs. 0.041) (Supplementary Table 5). The importance of structural features in predicting pathogenic AASs was also frequently emphasized in previous studies, since they can improve the prediction of nAASs in conserved/constrained regions and daAASs in regions with loose constraints, in addition to offering hints of pathogenic mechanisms [55][56][57].…”

MutTMPredictor: Robust and accurate cascade XGBoost classifier for prediction of mutations in transmembrane proteins

“…With the development and application of next-generation sequencing technology, a large amount of genetic mutation data has been detected, which can be utilized to study the correlation between human genetics and diseases [1] . The rapid identification of pathogenic genetic mutations can help understand the pathogenesis of diseases, which can also contribute to the early detection of disease and timely treatment [2] .…”