DNA methylation heterogeneity patterns in breast cancer cell lines

Abstract

Heterogeneous DNA methylation patterns are linked to tumor growth. In order to study DNA methylation heterogeneity patterns for breast cancer cell lines, we comparatively study four metrics: variance, I2 statistic, entropy, and methylation state. Using the categorical metric methylation state, we select the two most heterogeneous states to identify genes that directly affect tumor suppressor genes and high- or moderate-risk breast cancer genes. Utilizing the Gene Set Enrichment Analysis software and the ConsensusPath Database visualization tool, we generate integrated gene networks to study biological relations of heterogeneous genes. This analysis has allowed us to contribute 19 potential breast cancer biomarker genes to cancer databases by locating “hub genes” - heterogeneous genes of significant biological interactions, selected from numerous cancer modules. We have discovered a considerable relationship between these hub genes and heterogeneously methylated oncogenes. Our results have many implications for further heterogeneity analyses of methylation patterns and early detection of breast cancer susceptibility.