Clusters identification in binary genomic data: The alternative offered by scan statistics approach

Abstract

In many different research area, identification of clusters or regions showing an increment in event rate over a given study area is an important and interesting problem. Nowadays literature concerning scan statistics is quite broad and methods can be subdivided based on dimensional complexity of the study area, assumption on distribution generating the data under the null hypothesis and shape-dimension of the scanning window. The aim of this study is to adapt and apply this methodology to the genomics field taking into account for some peculiarities of these data and to compare its performance to existing method based on DBSCAN algorithm. © 2014 Springer International Publishing Switzerland.