A0209
Title: Mutual information for detecting multi-class biomarkers when integrating multiple omics studies
Authors: Jian Zou - Chongqing Medical University (China) [presenting]
Abstract: Biomarker detection is crucial in biomedical research. While integrating multiple omics studies improves the statistical power and robustness of results, current methods struggle with multi-class scenarios (e.g., various disease subtypes or treatments). Mutual information concordance analysis (MICA), a statistical framework for identifying biomarkers with consistent multi-class expression patterns across multiple studies, is introduced. MICA employs information theory to test and detect these biomarkers globally, followed by a post hoc analysis to pinpoint studies with concordant patterns. Extensive simulations and two real applications demonstrate the superior performance of the proposed method.
