Wei Pan

Mendel is a foundation model trained on human whole genome sequences to prioritize regulatory variants and improve gene expression prediction.

The paper introduces COWAS, a method to identify gene or protein pairs whose genetically regulated co-expression is associated with complex traits, finding interacting protein pairs linked to cholesterol, Alzheimer's, and Parkinson's disease.

DRAB identifies genes with context-specific genetic regulation by comparing models across tissues using bootstrapping and elastic net.

The paper infers a directed acyclic graph of phenotypes from GWAS summary statistics and applies it to cardiovascular proteins linked to Alzheimer's disease.

The paper uses deep learning to extract neuroimaging features from MRI data for Alzheimer's disease GWAS, identifying associated SNPs linked to neurodegenerative disorders.

The paper introduces a causal direction inference method for GWAS summary data that handles correlated pleiotropy and identifies bi-directional causal relationships between traits.

The paper introduces a new goodness-of-fit test, TEDE, to check model assumptions in Mendelian randomization and transcriptome-wide association studies, demonstrating its higher statistical power compared to existing methods.

This paper introduces new methods for inferring causal directions between traits using GWAS summary data, addressing limitations of Steiger's method by incorporating multiple SNPs and handling horizontal pleiotropy.

This paper evaluates the performance of standard TWAS versus 2SRI and error correction in TWAS, concluding that standard TWAS is effective for large samples and small effects.

This paper introduces a novel algorithm to integrate heterogeneous temporal scRNA-seq datasets and reconstruct lineage trajectories, identifying Ebf1 as a key transcription factor for mouse endothelial development.

The paper introduces a new algorithm for de novo discovery of mutated driver pathways in cancer, showing improved efficiency and scalability over existing methods.