Hagai Levi

PRANA: A Deep Learning Method for Adapting Polygenic Risk Scores to Diverse Ethnic Groups

PRANA adapts polygenic risk scores to non-European populations, improving predictive accuracy by 5%-20% in South Asian, East Asian, and Ashkenazi Jewish populations.

Mapping Genetic Risk Associations to Cellular Contexts via Deep Learning and Biological Ontologies

This paper maps GWAS risk variants to relevant cellular contexts using deep learning and biological ontologies to interpret genetic associations.

The predictive capacity of polygenic risk scores for disease risk is only moderately influenced by imputation panels tailored to the target population

This study evaluates how imputation panels affect polygenic risk score accuracy across different ethnic populations, finding only marginal improvements when panels match the target population.

The DOMINO web-server for active module identification analysis

The paper introduces a web-server for the DOMINO algorithm, which identifies active modules in gene networks and provides tools for analysis and interpretation.

DOMINO: a novel algorithm for network-based identification of active modules withreduced rate of false calls

The paper introduces DOMINO, a novel algorithm for identifying active modules in omics data, which outperforms existing methods by reducing false calls through permutation-based significance evaluation.