An Alignment-Free Distance Measure for Closely Related Genomes

Phylogeny reconstruction on a genome scale remains computationally challenging even for closely related organisms. Here we propose an alignment-free pairwise distance measure, K-r, for genomes separated by less than approximately 0.5 mismatches/nucleotide. We have implemented the computation of K-r based on enhanced suffix arrays in the program kr, which is freely available from guanine. evolbio.mpg.de/kr/. The software is applied to genomes obtained from three sets of taxa: 27 primate mitochondria, eight Staphylococcus agalactiae strains, and 12 Drosophila species. Subsequent clustering of the K-r values always recovers phylogenies that are similar or identical to the accepted branching order.