Abstract A53: Integrative data analysis of colorectal cancer reveals the frequently metastatic sample cluster

High-throughput biological technologies enabled to identify cancer related key molecules and subclasses. However, single experiment could only elucidate one feature of the tumor pathway. In order to identify the clusters and pathways significant for cancer progression, we performed integrated analysis using gene expression and DNA methylation profiles of colorectal cancer samples. Extracting core CpG sites from CpG island with our newly developed method, iClusterPlus, one of integrated analyses methodologies, was applied to expression array and methylation profiling data and 116 colorectal cancer samples were divided into various cluster size (2≤k≤11). We examined clusters (k=3 and k=6), and found that each cluster (k=3) clearly corresponds to three groups consisting of the combination of two clusters in cluster (k=6) with different prognosis. Expression profiles of two clusters in each group were significantly different, which suggests that aberrant DNA methylation is accumulated during the early stage of cancer, while alteration of gene expression profile is the late event of cancer progression. Furthermore, network-based analysis using expression profiles demonstrated that modules including hnRNP family was down-regulated in one of the cluster (k=6) which frequently harbored liver metastasis and poor prognosis. Integrated clustering analysis applied to multidimensional methods is available to identify disease characteristics such as survival rate and distant metastasis, which might be valuable for improving cancer therapy. Citation Format: Shingo Tsuji, Yutaka Midorikawa, Motoaki Seki, Tadatoshi Takayama, Yasuyuki Sugiyama, Hiroyuki Aburatani. Integrative data analysis of colorectal cancer reveals the frequently metastatic sample cluster. [abstract]. In: Proceedings of the AACR Precision Medicine Series: Drug Sensitivity and Resistance: Improving Cancer Therapy; Jun 18-21, 2014; Orlando, FL. Philadelphia (PA): AACR; Clin Cancer Res 2015;21(4 Suppl): Abstract nr A53.