Proteome-wide Prediction of the Functional Impact of Missense Variants with ProteoCast
Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Background
Dissecting the functional impact of genetic mutations is essential to advancing our understanding of genotype-phenotype relationships and identifying new therapeutic targets. Despite the progress in sequencing and CRISPR technologies, proteome-wide mutation effect prediction remains challenging. Here, we introduce ProteoCast, a scalable and interpretable computational method for proteome-wide classification of genetic variants and functional protein site identification. It relies solely on evolutionary information, leveraging protein sequence data across organisms.
Results
Using ProteoCast, we generated mutational landscapes for 22,169 Drosophila melanogaster protein isoforms, categorising over 293 million amino acid substitutions as functionally neutral, uncertain, or impactful. We validated our predictions with over 380 thousand natural polymorphisms observed in the Drosophila Genetic Reference Panel (DGRP) and Drosophila Evolution over Space and Time (DEST) datasets and with FlyBase’s developmentally lethal mutations. About 86% of known lethal mutations were classified as impactful or uncertain, versus only 13% and 18% of DGRP and DEST mutations. Moreover, we performed ProteoCast-guided genome editing experiments, providing a proof-of-concept of the validity of this strategy. Beyond variant effect prediction, ProteoCast detected evolutionary conservation signals in about one-third of 40.5K annotated post-translational modification sites and 83% of ∼90 known short linear motifs. These results support its usefulness for uncovering interaction and regulatory sites in unstructured protein regions.
Conclusions
Our results demonstrate ProteoCast applicability for model organisms, contributing to basic genetic research and translational studies. This work provides a publicly available dataset, user-friendly interactive web services, and a locally deployable pipeline tool for further research into gene function and mutation effects in any organism.
