Abstract: Detecting predictive biomarkers from multi-omics data is essential for precision medicine, to improve diagnostics of complex diseases and better treatments. This needs substantial experimental efforts that are made difficult by the heterogeneity of cell lines and huge cost. A practical solution is to build a computational model over the diverse omics data, including genomic, molecular, and environmental information. However, choosing informative and reliable data sources from among different data types is a challenging problem.
In this talk, I will present DIVERSE, a Bayesian importance-weighted matrix factorization framework to predict drug responses of cancer cell lines from newly combined data, which we assembled from five independent data sources. DIVERSE integrates the data sources systematically, in a step-wise manner to decide which information is useful to be incorporated and how significant the information is for the prediction task. The approach enables DIVERSE to predict values for entirely unseen drug response vectors to given cell lines by leveraging from previous experiments performed on similar drugs and cell lines.
Speakers: Betul Guvenc
Affiliation: Aalto University