Cong Liu

The paper develops a GPU-accelerated polygenic scoring method using an autonomous coding agent, achieving speedups on synthetic data but showing discrepancies on real data.

This study identifies manganese as a regulator of Akt activity independent of PI3K, showing it enhances Akt catalytic efficiency and interacts with insulin to modulate glucose homeostasis.

This paper fine-tunes Llama 2 for rare disease concept normalization using HPO data, achieving high accuracy with synonym inclusion and typo handling.

This paper evaluates methods to incorporate alternative polygenic risk scores into the BOADICEA breast cancer risk prediction model, finding that retrospective likelihood approaches provide more accurate estimates.

This study evaluates how ancestry, GWAS sample size, and LD reference panels affect BMI PRS performance across different age groups and ancestries.

This paper develops a genome-wide polygenic score for chronic kidney disease that incorporates APOL1 risk genotypes and shows consistent performance across four major continental ancestries.

The paper identifies pathogenic variants in arrhythmia genes in eMERGE-III participants and reclassifies variants of uncertain significance using functional studies.

The paper evaluates the generalizability of breast cancer polygenic risk scores across three ancestral groups using the eMERGE dataset, finding good performance in LatinX but not African ancestry women.

Phen2Gene prioritizes genes based on HPO terms for rare diseases, outperforming existing tools in speed and accuracy for clinical diagnosis.