Title: Meta-clustering with multi-level omics data for cancer subtype discovery
Authors: Yingying Wei - The Chinese University of Hong Kong (Hong Kong) [presenting]
Abstract: In traditional meta-analysis, we pool effect sizes across studies to improve statistical power. In meta-clustering, we want to conduct clustering jointly across studies. We propose a Bayesian hierarchical model that integrates diverse data types for clustering, accounts for the technical artifacts in individual studies, and handles cluster imbalance across studies. We apply the proposed methods to TCGA data and systematically identify subtypes for major human cancers.