Genomic and network analysis to study the origin of ovarian cancer

Abstract

Characterizing the origin of high-grade serous ovarian cancer has significant practical importance for advancing biological knowledge and improving clinical treatments. Rapid advances in molecular profiling technologies and machine learning based data analytics provide new opportunities to investigate this important question using data-driven approaches at the molecular and network levels. We now report novel analytic results in assessing the origin of high-grade serous ovarian carcinoma. Using genome-wide gene expression data and effective machine learning approaches, we design proper statistical significance tests and perform both genomic and network analyses to discriminate among three possible origins. The experimental results are consistent with recent scientific hypothesis and independent findings.