New approaches to identify cancer heterogeneity in DNA methylation studies using the lepage test and multinomial logistic regression

It is well known that cancer cells are diverse and heterogeneous among patients, and that this heterogeneity makes cancer diagnosis and cure difficult. Intra-tumoral heterogeneity has very recently become important because a small proportion of drug-resistant or tumor-initiating cells can ultimately determine a patient's outcome. In this study, we propose new approaches to use variance tests, including the Lepage test, to identify inter-patient cancer heterogeneity, as well as multinomial logistic regression to identify intra-patient cancer heterogeneity, i.e., different epiallele composition. We conduct experiments to show the performance of the proposed approaches.