Abstract:The clinical interpretation of missense variants is critically important in diagnostics due to their potential to cause mild-to-severe effects on phenotype by altering protein structure. Evaluating these variants is essential because they can significantly impact disease outcomes and patient management. Many computational predictors, known as in silico pathogenicity predictors (ISPPs), have been developed to support the assessment of variant pathogenicity. Despite the abundance of these ISPPs, their predictions often lack accuracy and consistency, primarily due to limited data availability and the presence of erroneous data. This inconsistency can lead to false positive or negative results in pathogenicity evaluation, highlighting the need for standardization. The necessity for reliable evaluation methods has driven the development of numerous ISPPs, each attempting to address different aspects of variant interpretation. However, the sheer number of ISPPs and their varied performances make it challenging to achieve consensus in predictions. Therefore, a comprehensive statistical approach to evaluate and integrate these predictors is essential to improve accuracy. Here, we present a comprehensive statistical analysis comparing 52 available ISPPs, which aims to enhance the precision of variant classification. Our work introduces the Variant Analysis with Multiple Pathogenicity Predictors-score (VAMPP-score), a novel statistical framework designed for the assessment of missense variants. The VAMPP-score leverages the best gene-ISPP matches based on ISPP accuracies, providing a combinatorial weighted score that improves missense variant interpretation. We chose to develop a statistical framework rather than creating a new ISPP to capitalize on the strengths of existing predictors and to address their limitations through an integrative approach. This approach not only improves the evaluation of missense variants but also offers a flexible statistical framework designed to identify and utilize the best-performing ISPPs. By enhancing the accuracy of genetic diagnostics, particularly in the reanalysis of rare and undiagnosed cases, our framework aims to improve patient outcomes and advance the field of genetic research.

Decoding Missense Variants by Incorporating Phase Separation via Machine Learning

Classification of MLH1 Missense VUS Using Protein Structure-Based Deep Learning-Ramachandran Plot-Molecular Dynamics Simulations Method

Discordance between a deep learning model and clinical-grade variant pathogenicity classification in a rare disease cohort

A New Era in Missense Variant Analysis: Statistical Insights and the Introduction of VAMPP-Score for Pathogenicity Assessment

DVA: Predicting the Functional Impact of Single Nucleotide Missense Variants

VariPred: Enhancing Pathogenicity Prediction of Missense Variants Using Protein Language Models

Enhancing missense variant pathogenicity prediction with protein language models using VariPred

Decoding molecular mechanisms for loss of function variants in the human proteome

IDRMutPred: Predicting Disease-Associated Germline Nonsynonymous Single Nucleotide Variants (nssnvs) in Intrinsically Disordered Regions

Enhancing Missense Variant Pathogenicity Prediction with MissenseNet: Integrating Structural Insights and ShuffleNet-Based Deep Learning Techniques

Predicting Pathogenic Single Nucleotide Variants Through a Comprehensive Analysis on Multiple Level Features

Comprehensive evaluation of AlphaMissense predictions by evidence quantification for variants of uncertain significance

Precise prediction of phase-separation key residues by machine learning

Clinical Utility of Protein Language Models in Resolution of Variants of Uncertain Significance in KCNQ1, KCNH2, and SCN5A Compared With Patch-Clamp Functional Characterization

A Bioinformatics Pipeline for the Identification of Disease-Causing Variants in Humans That Can Change Protein Structure.

Analysis of missense variants in the human genome reveals widespread gene-specific clustering and improves prediction of pathogenicity

Using deep mutational scanning to benchmark variant effect predictors and identify disease mutations

Understanding structure-guided variant effect predictions using 3D convolutional neural networks

Fine-tuning Protein Language Models with Deep Mutational Scanning improves Variant Effect Prediction

MVP predicts the pathogenicity of missense variants by deep learning

Integration of deep learning with Ramachandran plot molecular dynamics simulation for genetic variant classification