Genome-wide Profiling of DNA Methylation Patterns in Colon Cancers and its Utility in Biomarker Development for Early Detection and Prognosis

Sungwhan An, Chief Scientific Officer, Genomictree

To purchase a DVD containing all of the presentations featured at Microarray World Congress 2009 please go to the Select Biosciences website.

Launch presentation

Abstract

It has become evident that abnormal changes in DNA methylation (meDNA) occur during tumorigenesis. Accordingly, defining profiles of altered methylation pattern could be useful for developing methylation biomarkers for early detection and stratification of cancers. We established a method for scanning genome for methylation status of CpG sites utilizing the binding property of engineered methyl-DNA binding polypeptide to meDNA for isolation and CpG-microarray. With this, we fully described the relative genome-wide methylation patterns in primary tumor and paired non-tumor tissues of 12 patients with different stage of colon cancers (half of them developed disease-recurrence within 3 years) and 2 normal tissues of healthy individuals over common reference DNAs. Methylation patterns was discern enough to distinguish primary tumor from non-tumor tissues by unsupervised hierarchical clustering with entire data passed a given criteria of filtering. Furthermore, Welch ANOVA and Fisher’s Exact tests in the search of significant differences of DNA methylation sites associated with recurrence and disease free among tumors selected 50 methylation targets which yielded very good classification results when evaluated by hierarchical clustering and PCA. Validation by pyrosequencing-based methylation assay is undergoing and reproducibility test of 50-predictor of recurrence in independent study will be addressed in this presentation.

Launch presentation