Guan Lab

Department of Computational Medicine & Bioinformatics
Differentiating functions at the Isoform level

Networks at a high resolution

A lot of diseases, e.g. cancers, Alzheimers, are believed to be related to myths in alternative splicing that involves a lot of network biology. One goal of our research is to understand this complexity. We have developed a series of methods and resources for isoform-level data integration. Many of these techniques have lead to best-performing methods in community-based, benchmark studies.

Guan Lab wins DREAM Alzheimer's Disease Big Data Challenge

How soon will I lose my ability to talk, walk and think? What does my brain image say? Those are the first questions of Alzheimer's patients. We have now developed the best-performing tools to answer both questions. Among many of the most competitive teams around the world, both the prognosis prediction tool and the image analysis tool were the winning method in the DREAM Alzheimers Disease Big Data Challenge.
GuanLab wins 2014 DREAM RA Drug Responder challenge

GuanLab wins 2014 DREAM RA Drug Responder challenge

Some rheumatoid arthritis patients respond nicely and go through complete remission; others may even get worse. Unfortunately, prescription of drugs for RA patients is still primarily a trial-and-error process. The 2014 DREAM rheumatoid arthritis drug response challenge asks global expertise to create models to solve this problem. We applied our unique data mining approach for predicting drug response using genetic and clinical information, which was the winning method for all sub-challenges.
Guan Lab co-wins 2014 DREAM Broad DREAM Gene Essentiality Prediction Challenge

Guan Lab co-wins 2014 DREAM Broad DREAM Gene Essentiality Prediction Challenge

Why some cancer patients respond to chemo, while others don't? What biomarkers we can use to identify the most effective drug for each cancer patient? To address this question, we developed a rigorous method for identifying predictive biomarkers for cancer cell survivability. This is the winning algorithm in the 2014 DREAM Broad DREAM Gene Essentiality Prediction Challenge - sub challenge 2: selecting the most predictive biomarkers for different cancer cell lines.
Guan Lab co-wins 2013 DREAM8 Network Inference Challenge

Guan Lab co-wins 2013 DREAM8 Network Inference Challenge

How would cancer cells respond to therapy? How does the signaling network changes upon drug intervention? We developed an elegant model for predicting the rewiring of causal networks under unseen interventions. This model achieved the best aggregated performance (in accuracy) in the 2013 HPN-DREAM8 Breast Cancer Network Inference Sub-Challenge 2A and 2B: Time-course prediction. It was identified as 'most consistent performing' for this task, defined as ranked robustly first using random subsets for testing.