Characterization of Genomic MidRange Inhomogeneity

by Bechtel, Jason M.

Genomes encode many signals at multiple levels. They must therefore manifest inhomogeneity in their nucleotide composition. Inhomogeneity is identifiable from the shortrange where neighboring nucleotides influence the choice of base at a site, to the longrange, commonly known as isochores, where a particular base composition can span millions of nucleotides. This work explores the mostly overlooked phenomenon of genomic midrange inhomogeneity (MRI). MRI is defined as an interdependence between nucleotide choice and base composition over a distance of 301000 base pairs. Through the investigation of thousands of human genes, the results presented herein establish the existence of MRI throughout mammalian genomes, establish a link with strong local secondary structures and characterize the properties of MRI. A public computational resource has been created to support further study of genomic MRI.