Petko P. Fiziev

This study identifies host factors and cellular contexts governing anellovirus persistence through genomic, transcriptomic, and functional analyses.

This paper introduces SpliceAI2, a deep learning model that improves identification and quantification of splice-altering variants and their effects on transcripts and disease.

This study functionally characterizes 221,412 trait-associated variants using MPRA in 5 cell types, identifying 12,025 regulatory variants and revealing widespread epistasis and non-TF mechanisms.

FlexRV improves gene discovery for complex traits by modeling nonlinear relationships between variant annotations and phenotypes, outperforming existing methods in association detection and replication.

The paper benchmarks PrimateAI-3D and AlphaMissense on real-world cohorts, finding PrimateAI-3D performs better in rare-disease and biobank settings.

The paper uses primate genetic data to predict human variant pathogenicity with a deep learning classifier trained on 4.3 million common missense variants.

The paper identifies rare variant polygenic risk scores as more effective than common variant polygenic risk scores in identifying individuals with outlier phenotypes in common human diseases and complex traits.

CSREP summarizes chromatin state annotations within sample groups and identifies differential annotations across groups using probabilistic modeling and logistic regression.

The paper introduces ChromTime, a method to model spatial changes of chromatin mark peaks over time, identifying expanding, contracting, and stable regions linked to transcriptional changes.