Blumenstiel et al. 2002

Blumenstiel JP, Hartl Dl, Lozovsky ER. 2002. Patterns of insertion and deletion in contrasting chromatin domains. Molecular Biology and Evolution 19: 2211-2225.

These authors examined “dead-on-arrival” (DOA) transposable elements (TEs) in the genome of Drosophila melanogaster. They divided the genome into three categories for their analysis: heterochromatin, euchromatin of low recombination frequency, and euchromatin of high recombination frequency. They also calculated local rates of DNA gain or loss by small indels.

TEs were more abundant in heterochromatin than in either type of euchromatin, and more abundant in low-recombination euchromatin than in high-recombination euchromatin. Young TEs (more recent insertions) were more evenly distributed while older “mature” TEs were heavily biased towards heterochromatin. The average rate of DNA loss over the entire genome examined was 5.6 bp per nucleotide substition, a rate of spontaneous DNA loss much faster than found in mammals (Petrov and Hartl, 1998).

The authors propose four potential mechanisms to explain the observed distribution of TEs. The first two, that TEs insert preferentially in heterochromatin and that the rate of spontaneous deletion is lowest in heterochromatin, are each rejected. The examined TEs do not insert preferentially anywhere in the genome, as evidenced by the even distribution of the youngest TEs. Spontaneous deletion rates calculated in this study did not vary across regions of the genome, and were insensitive to local recombination rates. Small deletions in TEs should be favoured (and commonly detected) if the deletion rate varies in such an adaptive manner.

The remaining explanatory mechanisms both involve selection, either relaxed selection against TEs in heterochromatin relative to in euchromatin, or positive selection on TEs in heterochromatin for some reason. This analysis suggested the relaxed selection hypothesis is more likely, though it is neutral with respect to the relative contributions to that selection of avoiding ectopic recombination and avoidance of interference with gene function.

The authors suggest that while they have no evidence that small deletions are favoured any more in one region than another, they claim to have some evidence that large deletions, greater than 400 bp, may be favoured. However, this statement is based on a difference in mean deletion sizes between regions that is not statistically significant; they seem to be claiming a trend that does not exist! Perhaps I misread that section of the Discussion.