Xihong Lin

SPLENDID improves polygenic risk prediction across diverse populations by incorporating continuous genetic ancestry without requiring discrete ancestry labels.

A web atlas counts evidence from 24 types for 188,634,118 variant-gene pairs across 50 human tissues; 1,530,420 pairs pass a high-confidence cutoff, 87% of them tissue-specific.

The paper introduces SEEK-VFI, a method for selecting trajectory-relevant genes in single-cell analysis, which improves trajectory inference and downstream analyses.

SEEK-VEC improves topic modeling by combining multiple models through spectral ensembling, enhancing performance in identifying vocabulary and relationships, especially with weak signals.

This paper introduces graphREML, a likelihood-based method for heritability partitioning and enrichment analysis using GWAS summary statistics, showing greater power than S-LDSC.

STELLAR improves polygenic risk prediction by integrating rare and common variants using ensemble learning, enhancing accuracy and refining risk stratification.

The paper maps rare coding and noncoding variants to 1,342 diseases and biomarkers using whole genome sequencing data from 490,549 UK Biobank participants.

The paper introduces vcf2agds, a toolkit converting WGS data to aGDS format, enabling efficient analysis of UK Biobank data and identifying 480 genome-wide significant associations for total cholesterol.

This study identifies FABP4 as a host target that modulates SARS-CoV-2 infection, showing that its inhibition reduces viral replication and lung damage in infected hamsters.

The paper develops polygenic risk scores to stratify lung cancer and smoking cessation risk, showing significant odds ratios in diverse populations and planning clinical trials to evaluate interventions.

ALL-Sum improves polygenic risk prediction by combining efficient summary statistic-based regression with ensembling, outperforming existing methods in accuracy, speed, and memory usage.

This paper introduces SynSurr, a method that improves GWAS power for partially missing phenotypes by jointly analyzing the target phenotype and a synthetic surrogate predicted by machine learning, without requiring a correctly specified imputation model.

The paper identifies 83 rare lncRNA variant sets associated with blood lipid levels using WGS data from the TOPMed program and validates some in UK Biobank data.

The paper introduces CT-SLEB, a method for multi-ancestry polygenic prediction that improves performance in non-European populations using GWAS summary statistics and simulation studies.

The paper analyzes inflammation biomarkers using whole-genome sequencing from the TOPMed consortium, identifying both single-variant and rare variant associations with inflammatory traits.

The paper introduces HEELS, a new estimator for local heritability using GWAS summary statistics and LD matrix, achieving high statistical efficiency comparable to REML.

FAVOR provides an online resource and tool for functional annotation of variants from whole genome sequencing studies to aid in disease gene prioritization.

The paper performs whole genome sequence analysis of plasma lipid levels in 66,329 individuals, identifying new lipid-associated alleles, including rare non-coding variants.

The paper performs whole genome sequence association analysis of fasting glucose and insulin levels in diverse cohorts, identifying 22 significant variants across 16 gene regions, including rare variants.

The paper introduces a cloud-based workflow for rare variant analysis in whole genome sequencing studies, enabling scalable and reproducible analysis with functional annotations.

The paper estimates heritability for height and BMI using whole-genome sequence data and finds that rare variants contribute significantly to the missing heritability.

The paper introduces a method to improve eQTL inference in undersampled tissues by jointly modeling a target outcome and a correlated surrogate outcome.

The paper introduces MACIE, a framework for predicting functional variants by integrating diverse annotations, and applies it to fine-mapping using GWAS data.

The paper reports whole-genome association analyses of sleep-disordered breathing traits in the NHLBI TOPMed program, identifying rare and common variant associations.

AFA estimates ancestral allele frequencies in admixed populations with high accuracy, showing strong correlation with gnomAD frequencies in the Hispanic Community Health Study/Study of Latinos.

The paper develops a new statistical test, DACT, to detect mediation effects in genome-wide epigenetic studies and applies it to identify DNA methylation sites mediating smoking's effect on lung function.

Over 400 million variants, 97% below 1% frequency, emerge from deep whole-genome sequencing of 53,831 diverse TOPMed participants; TOPMed haplotypes enable imputation down to ~0.01% frequency.

APEX is a toolkit for molecular QTL mapping and meta-analysis that improves scalability and accuracy for large datasets and summary statistics.

The paper introduces a frailty model for genome-wide survival analysis in large biobanks, handling population structure and relatedness, and demonstrates its performance on real-world datasets.