Horizontal gene transfer, genome innovation and evolution

Top Cited Papers
Publisher Website
Google Scholar
Abstract
Tree-like binary schemes for taxonomic classification have an illustrious history in evolutionary biology, but they do not provide a complete representation of life's history, especially for prokaryotes. Some genes within an organism have tree-like histories that differ from the histories of other genes within the same organism, owing to horizontal gene transfer (HGT). Although all types of genes can be susceptible to horizontal transfer, different types of genes and groups of organisms vary in their propensity for HGT. Phylogenetic analysis of a concatenation of genes with differing gene histories has the potential to create artifactual histories that reflect neither the history of the organism nor the history of the gene. Methods of phylogenetic reconstruction that do not insist on tree-like phylogenies but explicitly allow for reticulation events promise to yield a more realistic reconstruction of much of life's history. Phylogenetic reconstruction methodology — and therefore the accurate detection of historical HGT events — can benefit from a better understanding of phylogenetic signal as it relates to rates of change of characters and taxonomic sampling. The abundance and atypical composition of genes that are transient in prokaryotic genomes such as that of Escherichia coli lead one to wonder whether such genes (and horizontally transferred genes in general) are deleterious, selected for or neutral on transfer. Recent studies seem to indicate extensive gene swapping and few selective sweeps, arguing that many transferred genes might be nearly neutral in selective effect. Quantitative methods for analysing the frequency and nature of HGT events in microbial communities are needed. Quantitative descriptions yield precise predictions that can be tested statistically, and will help to resolve disputes about the frequency and importance of HGT to microbial evolution by improving clarity of expression, as well as enforcing statistical decision criteria.