Xiaoyu Liang

This paper improves polygenic score prediction for underrepresented groups using transfer learning methods implemented in the GPTL R-package, showing better performance than traditional approaches.

The paper identifies CpG sites and pathways associated with gestational age and epigenetic age acceleration, and links perinatal exposures to these epigenetic changes.

This study constructs a bipartite Genotype and Phenotype Network to analyze genetic associations and improve phenome-wide association studies by identifying pleiotropic loci and controlling false discovery rates.

This study shows hemin induces HO-1 in macrophages, inhibiting STING pathway activation and promoting endometriosis development.

The paper introduces IntQuery, a deep learning model that predicts LSI off-target activity genome-wide using data from Cryptic-seq.

The paper develops and validates methods to detect off-target recombination by Large Serine Integrases, finding higher sensitivity than whole genome sequencing.

This study identifies a shared role for mec-2 and Stoml3 in olfactory behavior across C. elegans and mice, showing that they are required for odor discrimination but not odor detection.

The paper proposes gene selection methods that integrate genetic networks into case-control association studies, showing improved true positive rates for identifying rheumatoid arthritis-related genes.

This paper introduces HCLC-FC, a multivariate method for PheWAS that improves FDR control and identifies significant SNPs in large-scale EMR data.

This study develops a monocyte-based DNA methylation clock (MonoDNAmAge) and shows its nonlinear relationship with alcohol consumption across three cohorts.

The paper predicts cellular positions in the Drosophila embryo using scRNA-seq data by benchmarking methods against spatially annotated genes.

This study identifies 294 genes affected by acute THC administration in human immune cells, including 69 common and 225 cell-type specific genes.

The paper identifies a DNA methylation signature linked to phosphatidylethanol (PEth) that better predicts hazardous alcohol consumption than self-reported data in two populations.