Horizontal gene transfer: pitfalls and promises.

Abstract

J. PETER GOGARTEN, RYAN D. MURPHEY, AND LORRAINE OLENDZENSKI Department of Molecular and Cell Biology, University of Connecticut, Storrs, Connecticut 06269-3044 Phylogenetic reconstruction from protein or nucleic acid sequence families provides information on the evolution of individual genes. In contrast to the assumed bifurcating, tree-like evolution of genes, organismal evolution is char- acterized by the exchange of genetic information between organisms and even by the fusion of formerly independent lines of descent (1). The invocation of horizontal gene transfer events is often regarded as a last-ditch attempt by systematists to reconcile conflicting phylogenies con- structed from different markers. In general, however, organ- ismal evolution is clearly visible as the majority consensus of a number of molecular phylogenies, and transfer events can be recognized in phylogenies constructed from one or several markers whose topologies deviate from that of the consensus. Often it is difficult to decide whether conflicts between molecular phylogenies are due to actual events in evolution (horizontal gene transfer or gene duplications [see 2]), or due to artifacts generated during phylogenetic reconstruc- tion. For example, investigation of the maximum likelihood landscape of 18s rRNA and V-ATPase A-subunit phylog- enies suggests that the grouping of microsporidia either with the fungi (3,4,5) or close to the root of the archaeal domain (e.g., 6, 7, 8) probably represents an artifact in the 18s rRNA data analyses and not another case of horizontal transfer. Although the recognition of horizontal transfer as a major factor in prokaryotic evolution (e.g., 9, 10, 11, 12) certainly complicates the interpretation of molecular phylogenies, it also allows synchronization of different parts of the univer- sal tree of life, and thus might provide the key to the detection of periods of rapid substitutions. Two examples