Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
September 16, 2025Open Access

Disease-specific variant pathogenicity prediction using multimodal biomedical language models

View Full Paper
Ask AI
Bookmark
Share

Authors

YLYilin LiuDCD.N. CooperHYHaiyuan Yu

Discussion

Loading...

Member takes

Overview

A new deep learning method predicts variant effects in disease-specific contexts, suggesting improved clinical interpretation.

Key Points

  • DIVA accurately predicts disease-specific pathogenicity of missense variants, enhancing clinical utility for diagnosis.
  • The integration of protein sequences and disease annotations within a contrastive learning framework significantly improves predictions.
  • Benchmark results show that DIVA outperforms existing baseline methods in classifying variant deleteriousness.
  • Utilizing AlphaMissense scores within DIVA boosts the accuracy of deleterious variant classification significantly.

Cite This Study

Liu et al. (2025) studied this question.

synapsesocial.com/papers/68d454bb31b076d99fa59e95https://doi.org/10.1101/2025.09.09.675184
View Full Paper
Ask AI
Bookmark
Share