Abstract 2995: Identifying subtypes of colorectal cancer tumors and cell lines with treatment relevance using iterative clustering (iNMF)

Abstract

Genetic and epigenetic features such as KRAS mutation status have been used to define colorectal cancer (CRC) subtypes and to take treatment decisions. In order to develop new targeted drugs, however, it is necessary to gain a better understanding of the molecular differences of CRC subtypes. We developed a new unsupervised approach for stratifying tumor samples using genome-wide mRNA expression data. Our method is based on the iterative application of non-negative matrix factorization (iNMF) on randomly selected sets of genes. In a sample set consisting of 63 CRC tumors, we identified two dominant subtypes. These subtypes were highly concordant with the epithelial-mesenchymal-transition (EMT) gene expression signatures consistent with recent publications. However further stratification of the tumor samples revealed five subtypes. These subtypes exhibit distinct differences, most notably differential activation of specific signaling pathways. Importantly, assessment of this method and derived subtype gene signatures stratified several independent, published datasets, suggesting that the signatures capture disease-relevant intrinsic features of CRC. Furthermore, application of the gene signatures to expression data obtained from three independent colon cell line panels revealed that the tumor subtypes were represented in all these different panels. Additional integration of pharmacological response data allowed us to identify several targeted compounds showing differential response across the subtypes. The CRC stratification obtained with our new method, iNMF, offers valuable insight into the differences between CRC subtypes at a functional level. Most importantly, it captures features of the disease that are highly relevant for the development of new targeted drugs in defined CRC patient sub-populations. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 2995. doi:1538-7445.AM2012-2995