Jiacheng Miao

The Virtual Biotech: A Multi-Agent AI Framework for Therapeutic Discovery and Development

The Virtual Biotech uses AI agents to integrate multi-omic and clinical data for drug discovery, identifying genomic features linked to trial success and proposing therapeutic strategies.

Polygenic prediction of treatment efficacy with causal transfer learning

This paper introduces a statistical framework to identify genetically driven heterogeneous treatment effects using causal transfer learning, applied to RCTs to find key determinants of treatment efficacy.

Genetic basis of partner choice

This study performs the first GWAS on a latent partner choice index using 206,617 individuals, identifying genetic correlates and solutions to reduce assortative mating biases in genetic studies.

Controlling for polygenic genetic confounding in epidemiologic association studies

This study introduces PENGUIN, a framework to control genetic confounding in epidemiologic studies, showing it effectively removes confounding in educational attainment associations.

Optimizing and benchmarking polygenic risk scores with GWAS summary statistics

This paper introduces a statistical framework to optimize and benchmark polygenic risk scores using GWAS summary statistics, showing improved performance over existing methods.

Valid inference for machine learning-assisted GWAS

This study identifies risks of false positives in ML-assisted GWAS and introduces POP-GWAS, a framework that provides valid inference for GWAS on ML-imputed phenotypes, applied to BMD data to find 89 novel loci.

Pervasive biases in proxy GWAS based on parental history of Alzheimer's disease

This study identifies biases in proxy GWAS using parental Alzheimer's history, showing misleading genetic correlations and highlighting design issues in biobank cohorts.

Neuropathology-based APOE genetic risk score better quantifies Alzheimer's risk

The paper develops an APOE genetic risk score based on neuropathology that better explains CSF biomarker variance than traditional APOE ε4 metrics.

Quantifying portable genetic effects and improving cross-ancestry genetic prediction with GWAS summary statistics

This paper introduces X-Wing, a framework that improves genetic prediction across populations using GWAS summary statistics by identifying portable genetic effects.

Reimagining Gene-Environment Interaction Analysis for Human Complex Traits

This study introduces PIGEON, a statistical framework for gene-environment interaction analysis using summary statistics, showing its superiority in multiple analytic settings.

A quantile integral linear model to quantify genetic effects on phenotypic variability

This paper introduces QUAIL, a model for detecting genetic variants affecting trait variability, and finds 11 novel BMI vQTL with interactions with physical activity and sedentary behavior.