Assessment of compositional heterogeneity within and between eukaryotic genomes.

Using large amounts of long genomic sequences, we studied the compositional patterns of eukaryotic genomes. We developed a simple measure, the compositional heterogeneity (or variability) index, to compare the differences in compositional heterogeneity between long genomic sequences. The index measures the average difference in GC content between two adjacent windows normalized by the standard error expected under the assumption of random distribution of nucleotides in a window. We report the following findings: (1) The extent of the compositional heterogeneity in a genomic sequence strongly correlates with its GC content in all multicellular eukaryotes studied regardless of genome size. (2) The human genome appears to be highly compositionally heterogeneous both within and between individual chromosomes; the heterogeneity goes much beyond the predictions of the isochore model. (3) All genomes of multicellular eukaryotes examined in this study are compositionally heterogeneous, although they also contain compositionally uniform segments, or isochores. (4) The true uniqueness of the human (or mammalian) genome is the presence of very high GC regions, which exhibit unusually high compositional heterogeneity and contain few long homogeneous segments (isochores). In general, GC-poor isochores tend to be longer than GC-rich ones. These findings indicate that the genomes of multicellular organisms are much more heterogeneous in nucleotide composition than depicted by the isochore model and so lead to a looser definition of isochores.