Lifan Liang

The paper identifies context-specific genetic effects in neuronal activation using single-cell multiomics, revealing NPD risk variants and genes active during stimulation.

This study develops M-cTWAS to integrate multi-omics and multi-context QTL data with GWAS, improving the discovery of risk genes for complex traits.

MIRAGE improves rare variant association analysis by incorporating variant functional information, outperforming existing methods in simulations and autism exome-sequencing data.

The paper introduces peak2vec, a method that uses a modified convolutional neural network to infer transcriptional regulation from ATAC-seq data, showing improved performance in distinguishing TFs compared to conventional methods.

This study learns gene embeddings from biomedical literature to uncover functional relationships and improve identification of cancer driver genes and protein complexes.

This study introduces ORN, a model that infers pathway activities and functional relationships in cancer using mutation and gene expression data, identifying survival-related pathways and novel mutation mechanisms.