Computational & Statistical Systems Biology Laboratory
@ NUSmed, National University of Singapore
We develop computational and statistical solutions for the analysis of high-dimensional molecular data in various contexts of biomedical research. These tools are motivated by a wide range of research topics, including
Integration of multi-omics data with biologically sensible molecular feature selection and robust outcome prediction
Development of workflows for large-scale untargeted and targeted mass spectrometry (MS) data extraction in proteomics and metabolomics
Effect size-driven analysis approaches for protein modifications, interactions, and other applications of proteomics
Using single-cell data as a prior information to make cell population-level inference from bulk -omics data
Interactive data visualization and exploration of high-dimensional data
Clinical application of sequencing and mass spectrometry-based omics technologies
To see examples of our work, check out the CSSB LAB GitHub page!
5 July 2021: ParProx, a state-of-the-art implementation for time-to-event and classification analysis using overlapping group lasso regression, is now available at Briefings in Bioinformatics. The tool enables scalable regression modelling of ultrahigh-dimensional omics data via parallel and distributed computing proximal gradient method, where the variable groups serve as an interface to guide the model selection with biological prior. Check out the GitHub repository. In collaboration with a wonderful group of computational statistics experts at Seoul National University (https://sites.google.com/site/johannwon/, https://kose-y.github.io/).
17 April 2021: New article delineating a DIA-based workflow with application to plasma metabolomics study in diabetic nephropathy is in Metabolites. We showcase MetaboKit tool for the processing of DDA and DIA-MS data, followed by the inferred partial correlation network-based multivariate classifier iOmicsPASS2+ (coming soon).
17 April 2021: Web-based multi-omics visualization tool multiSLIDE is published in Nature Communications. The source code available at GitHub repository: https://github.com/soumitag/multiSLIDE/. In combination with single omics data visualization SLIDE, multiSLIDE enables visualization of multiple omics data sets from genomics to metabolomics, revealing intricate relationship in molecular variations across different omics data sets.
20 January 2021: iOmicsPASS+, an extension of the first method paper at npj Systems Biology and Applications, is available at GitHub. The new software is equipped with new functionalities to infer a confounding-free partial correlation network from single- and multi-omics datasets and to perform supervised subnetwork analysis thereafter. Applicable to any omics data for cell/tissue analysis as well as circulating marker analysis.
Principal Investigator: Hyungwon Choi
Computational & Statistical Systems Biology Lab
Department of Medicine, Yong Loo Lin School of Medicine
National University of Singapore
Email: hwchoi [ at ] nus.edu.sg
HC is an associate editor at Molecular Omics -- feel free to discuss submission of your manuscript.