Hyungwon Choi

fastACCORD enables efficient ultrahigh-dimensional partial correlation modeling for multi-omic data integration, revealing cancer-specific methylation-expression dependencies.

The paper introduces DABEST 2.0, an estimation framework for multi-group comparisons in experimental science, focusing on effect-size quantification instead of null-hypothesis significance testing.

This paper develops a high-quality, scalable workflow for whole-cell single-cell RNA sequencing of adult cardiomyocytes and non-myocytes using in-cell ligation and split-pool barcoding, improving data quality and throughput.

This study identifies molecular changes in early MASLD progression in obesity through transcriptomic and metabolomic analyses, revealing distinct metabolic pathways and novel gene targets for liver fibrosis.

MRMhub automates processing of large-scale targeted metabolomics data, offering scalability and quality control tools, and is validated on lipidomics datasets.

This study quantifies site-specific ribosome ubiquitination under oxidative stress and identifies stress-specific ubiquitin modification patterns.

The paper studies two mutations in the zebrafish Kcng4b gene, showing how they affect Kv channel function and lead to distinct developmental defects in ear development.

The paper re-mines serum proteomics data to identify post-translational modifications associated with Zika and dengue infections, highlighting specific modifications linked to each virus.

The paper identifies molecular signatures of proteostasis differences between ALS-resistant and -sensitive motor neurons using transcriptome and proteome analysis.

ParProx enables scalable multivariable regression analysis of ultrahigh-dimensional omics data using group lasso with biological priors, producing interpretable models.

multiSLIDE is a web tool for visualizing interconnected molecular features in multi-omics data through keyword searches and customizable heatmaps.

The paper introduces a protein-centric method to aggregate exome variants for association analysis with clinical outcomes, identifying sequence units predictive of survival that conventional gene-level analysis misses.

This study uses high-resolution proteomics to identify 26 differentially expressed proteins that can predict Zika virus infection with 72% accuracy using a machine learning model.

The paper advocates for using estimation plots over p-values for data analysis and presents the DABEST tool for creating these plots.

iOmicsPASS integrates multi-omics data over biological networks to predict phenotypes and identify predictive subnetworks, improving prediction quality compared to methods without relational information.

PTMscape predicts post-translational modification sites across the proteome and identifies hotspots of modification crosstalk using a unified framework.

The paper quantifies gene expression regulation at multiple layers in response to ER stress in cervical cancer cells over time, identifying shared and stress-specific regulatory events.

This study uses a new statistical tool to analyze time-series mRNA and protein data, finding that both levels of regulation are important but have distinct dynamics in response to misfolding stress.