Fyodor A. Kondrashov

The paper evaluates zero-shot models' ability to predict epistatic effects on protein properties and finds they fail to predict strongly epistatic mutation combinations.

The paper identifies specific amino acid residues and N-glycosylation patterns in Bouncer proteins that mediate species-specific sperm-egg compatibility between medaka and zebrafish.

This study assesses AlphaFold's ability to predict the impact of single mutations on protein stability and function, finding weak or no correlation with experimental data.

The paper investigates the effect of an inherited POLD1 variant on mutation rates in germline and somatic cells, showing a recessive effect on proofreading deficiency in cancer.

The paper characterizes fitness landscapes of fluorescent proteins and finds that flat landscapes are less suitable for machine learning-based protein design.

The paper introduces a regressive alignment algorithm that starts with the most dissimilar sequences, producing more accurate alignments for large datasets.

The paper experimentally assays a fitness landscape on a macroevolutionary scale by analyzing the effects of amino acid substitutions in yeast His3 orthologs.

The paper predicts SNPs affecting drug-protein binding affinities using genomic and drug interaction data, estimating that individuals carry ~6 SNPs impacting ~5 FDA-approved drugs.