The Huang laboratory’s main research focus is translational pharmacogenomics with particular interest in the pharmacogenomics of anti-cancer agents. By systematically evaluating human genome and its relationships to drug response and toxicity, our goal is to develop clinically useful models that predict risks for adverse drug reactions and non-response prior to administration of chemotherapy. With her broad training background, Dr. Huang assembles and leads a multi-disciplinary team that consists of computational biologist, geneticist, physician, molecular biologist and biostatistician to tackle a series of serious problems in cancer research. These include the lack of mechanistic understanding of genomic regulation of cancer phenotypes; the lack of reproducible predictive biomarkers for cancer therapeutic agents; and the lack of effective treatment for many hard to treat cancers.
Bridging pre-clinical drug screening with patient genomic profiles in order to accurately predict patient response to therapy.
By leveraging new data and improved methods, the Huang lab has shown that gene expression based models derived from very large panels of cancer cell lines could directly inform clinical response to drugs. Furthermore, using an HDAC inhibitor as an example, they elucidated the polygenic architecture of drug response by comparing and contrasting the predicted power derived from a single gene, a pathway, or all genes in the genome. In this case, they clearly demonstrated the superior power of prediction when applying a machine learning method in deriving polygenic predictors of drug response. More recently, they described a novel statistical approach that improves the success rate of biomarker discovery by conditioning on shared drug sensitivity features among many drugs. Their work showed that despite a controversial history, drug screening in large panels of cell lines is more useful than was previously appreciated. Recently, they integrated data from large clinical cancer studies (e.g. The Cancer Genome Atlas (TCGA)) with data from pre-clinical disease models to impute drug sensitivity in patient data sets. These newly imputed drug sensitivity phenotypes enable large sequencing studies such as TCGA to be effectively used for pharmacogenomics discovery, which has previously been a severe limitation of these data.
PGRN Hub & the Featured Investigator