Because of the increase in paralogs through duplication, the rice genome may have undergone potentially many gene-conversion and unequal-crossover events in its evolution.
Given that the donor sequence remained unchanged in our HR events, and that many bidirectional gene conversion tracts were very long, these data suggest repair by two-ended SDSA.
Because many factors, including gene conversion between paralogous genes, could create highly polymorphic genes, we first clarified the pattern of molecular evolution of the CYP1A subfamily.
These nine gene families show structures that seem to promote gene conversion: many of them are included in large palindromes and others in inverted repeats and tandem arrays, which could explain the very high rate observed (Rozen et al. 2003).
Intriguingly, when we use the same approach with the synonymous substitutions in the RP genes, we find many fewer cases where gene conversion needs to be invoked: only 10 of 29 duplicate pairs show any evidence of gene conversion at the synonymous sites.
Many computational methods have been devised for detecting gene conversion (e.g. [ 12, 14- 16]).
It has been demonstrated that gene conversion is GC-biased in many eukaryotes, and in such cases, the probability that a GC allele is passed on to the next generation through gene conversion is higher than that of an AT allele.
It appears that many of these polymorphisms are the result of gene conversion introducing variation from the GOT1p1 copy to the GOT1p2 copy as can be seen by examining patterns of divergence between alleles.
For example, as described in the introduction, it has been demonstrated that gene conversion is GC-biased in many eukaryotes including humans and other mammals (Duret and Galtier 2009).
Gene conversion is a plausible mechanism for sequence homogenization among many tandem gene repeats and it is reported that its frequency is lower between separated gene copies [ 25- 27].
