Jonathan Frazer

The paper blends physics-based and inverse folding models to separate variant effects on protein stability and function, revealing distinct signatures for gain-of-function and loss-of-function variants.

This paper introduces popEVE, a deep generative model for proteome-wide disease variant prediction, identifying 442 genes in a developmental disorder cohort, including 123 novel candidates.

This paper improves variant effect prediction by introducing LFB, which averages model scores across sequences under similar selective pressures, enhancing performance across model sizes and datasets.

LOL-EVE predicts promoter variant effects using a transformer model trained on mammalian promoter sequences, showing improved performance in indel prediction and haplotype analysis.

ProteinGym unifies over 250 deep mutational scans and ~65k clinical variants to benchmark over 70 models in zero-shot and supervised settings; TranceptEVE ranks best overall.

MaveDB v2 curates over three million variant effects from MAVE studies to facilitate analysis and dissemination of functional assay data.

The paper predicts the pathogenicity of 11 million genetic variants across 1,081 disease genes using a deep learning model that leverages evolutionary data instead of clinical labels.