Shihong Mao

This paper introduces MonsterLM, a method for estimating gene-by-environment interactions on large datasets, finding significant interaction variance and improved polygenic score prediction.

This study develops polygenic risk scores for gestational diabetes in South Asian women using multi-ethnic summary statistics and validates them in UK Biobank data, finding moderate heritability.

The paper investigates X-chromosome variance heterogeneity in quantitative traits using generalized Levene's test and sex-stratified analysis, finding possible polygenic structure for human height variance.

The paper introduces a method to estimate regional polygenic correlation between complex traits and finds heterogeneity in shared heritability across genomic regions.

This paper introduces GraBLD, a machine-learning heuristic that improves polygenic risk scores for height, BMI, and diabetes using summary statistics from large-scale genome-wide studies.