Topological analysis of gene expression arrays identifies high risk molecular subtypes in breast cancer

摘要: Genomic technologies measure thousands of molecular signals with the goal understanding complex biological processes. In cancer these have been used to characterize disease subtypes, signaling pathways and identify subsets patients specific prognosis. However for any type are so vast that novel mathematical approaches required further analyses. Persistent computational homology provide a new method our previous work we presented homology-based supervised classification copy number aberrations from comparative genomic hybridization arrays. this first propose theoretical framework second extend analysis gene expression data. We analyze published breast data set find can distinguish most, but not all, different subtypes. This result suggests relationships between genes, captured by algorithm, help topological methods be clustering profiles.