Logit analysis of genetic data

Abstract

A new framework of genetic sequence statistical analysis based on generalized logit model is introduced. Logit analysis is applied to assess the dependence structure (interactions) between DNA nucleotides and to test hypothesis about Markov order of these dependencies. The procedure proposed seeks the non-coding subsequences which are homogeneous but yet non-Markov. It has been shown, that even homogeneous DNA regions can not be treated as the first order Markov sequences.