Looking at the tree of life, descent with modification is an obvious theme, where genes are passed on through 'vertical' lines of ancestry. It so happens though that genes can jump from one lineage to another, by a process called 'horizontal gene transfer' (HGT). Naked DNA uptake (transformation), viruses (transduction), and plasmids (conjugation) are the mechanisms by which the genetic units of heredity need not be inherited in the usual sense. HGT appears to blur the boundaries of what a species is, particularly for the bacterial domain of life. So the study published by Rotem Sorek, Edward Rubin et al.1 on the determination of barriers to HGT is interesting from a couple different perspectives.
[Sorek and coworkers] studied the attempted movement of 246,045 genes from 79 prokaryotic genomes into Escherichia coli and identified genes that consistently fail to transfer. We studied the mechanisms underlying transfer inhibition by placing coding regions from different species under the control of inducible promoters.
The resulting data can be found in the Science article, and the authors conclude that:
Our results suggest that there are universal gene-transfer barriers, regardless of whether transfer occurs among closely or distantly related microorganisms, and that these barriers may be associated with toxicity of the transferred gene to the host… Our observation that many unclonable genes are universally found as a single copy (never duplicated in any sequenced bacteria) suggests that the increased expression of these genes inhibits growth in a wide range of bacteria.
And that makes a lot of sense – we would expect that any genetic element successfully transferred in the wild would have a useful function, and that neutral or toxic genes would be selected against. In other words, horizontal gene transfer events would necessarily have to obey the same rules as their vertical brethren.
So, is there a subset of genes that is inaccessible to HGT, and therefore useful in reconstructing a "tree" of life? As McInerney and Pisani note in their commentary2:
Evolutionary biologist W. Ford Doolittle stated that if horizontal gene transfer was not limited to special categories of genes, then no system of classification could be considered natural.