Title: Nonparametric homogeneity tests and multiple change-point estimation for analyzing large Hi-C data matrices

Abstract: We propose a novel nonparametric approach for estimating the location of block boundaries (change-points) of non-overlapping blocks in a random symmetric matrix which consists of random variables having their distribution changing from one block to the other. Our method is based on a nonparametric two-sample homogeneity test for matrices that we extend to the more general case of several groups. We first provide some theoretical results for the two associated test statistics and we explain how to derive change-point location estimators. Then, some numerical experiments are given in order to support our claims. Finally, our approach is applied to Hi-C data which are used in molecular biology for better understanding the influence of the chromosomal conformation on the cells functioning.