The Effects of Sequence Removal on DNA-hybridization Estimates of Distance, Phylogeny, and Rates of Evolution

Abstract

The robustness of DNA-hybridization estimates of distance, phylogeny, and rates of evolution was examined via a technique called absorption. Single-copy nuclear DNAs of a study group consisting of four heron species and an ibis were hybridized with DNA from absorber species (either a swallow or stork) to remove shared sequences before they were hybridized among themselves. The amount and kind of DNA that was removed by absorption depended largely upon whether the absorber species was closely or distantly related to the study group and whether it was an outgroup or ingroup member. Absorption with outgroups, which removed between 25% and 80% of the genome, did not alter branching or evolutionary rate patterns but did increase distances slightly and reciprocal measurement asymmetry and tree-fitting error markedly. Increases in distances and homoplasy are expected when conservative sequences have been removed and mainly rapidly evolving sequences are left to hybridize. Absorption with ingroup DNA (i.e., ibis absorbed with stork) also did not affect branching, but it yielded other contrasting results. Genetic distances between ibis and herons were shortened rather than length? ened, and the pattern of rate variation known to occur among heron lineages was negated. The DNA sequences remaining after ingroup absorption appeared to be paraphyletically conserved; they were common to ibis and herons but not to ibis and its sister taxon, the storks. Overall, the effect of our sequence removal experiments was to increase site- and sequence-level noise in the DNA hybridization data without obliterating the phylogenetic signal. (Absorption; Ciconii- formes; conserved sequences; DNA hybridization; evolutionary rates; genome size.)