Identification and measurement of neigbor dependent nucleotide substitution processes

Abstract

The presence of different neighbor dependent substitution processes generates specific patterns of dinucleotide frequencies in all organisms. Based on a general framework of how to include such processes into more realistic models of nucleotide substitutions we develop a method that is able to identify such processes, measure their strength, and judge their importance to be included into the modeling. Starting from a model for neighbor independent nucleotide substitution we successively add neighbor dependent substitution processes in the order of their ability to increase the likelihood of the model describing the data. The analysis of neighbor dependent nucleotide substitutions in human, zebrafish and fruit fly is presented.