Non-alignment comparison of human and high primate genomes

TL;DR

This study uses compositional spectra analysis to compare human and primate genomes, revealing conserved synteny and phylogenetic signals in both coding and noncoding regions through k-mer based similarity measures.

Contribution

It introduces a novel approach combining compositional spectra and k-mer analysis to detect conserved genomic regions beyond gene anchors, including noncoding and repetitive DNA.

Findings

High correspondence in whole-genome comparisons

Revealed phylogenetic signals in noncoding regions

Combining GC content with k-mer abundances improves similarity detection

Abstract

Compositional spectra (CS) analysis based on k-mer scoring of DNA sequences was employed in this study for dot-plot comparison of human and primate genomes. The detection of extended conserved synteny regions was based on continuous fuzzy similarity rather than on chains of discrete anchors (genes or highly conserved noncoding elements). In addition to the high correspondence found in the comparisons of whole-genome sequences, a good similarity was also found after masking gene sequences, indicating that CS analysis manages to reveal phylogenetic signal in the organization of noncoding part of the genome sequences, including repetitive DNA and the genome "dark matter". Obviously, the possibility to reveal parallel ordering depends on the signal of common ancestor sequence organization varying locally along the corresponding segments of the compared genomes. We explored two sources contributing to this signal: sequence composition (GC content) and sequence organization (abundances of k-mers in the usual A,T,G,C or purine-pyrimidine alphabets). Whole-genome comparisons based on GC distribution along the analyzed sequences indeed gives reasonable results, but combining it with k-mer abundances dramatically improves the ordering quality, indicating that compositional and organizational heterogeneity comprise complementary sources of information on evolutionary conserved similarity of genome sequences.