Hi
I've been using SLAC, FEL and REL to identify codons under positive and negative selection in genes of CMV. So far, I've just been analysing each gene individually. However, in the virus several genes are overlapping (mostly on the opposite strand). Does this need to be taken into account, and if so does anyone have any ideas of how to do it?

Hi Lee,

Yes, this should be taken into account, but there is no proper way (other than complex MCMC treatment in a limited setting, which HyPhy can't do) to do so.

Do you see islands of low variability in overlapping regions?

Sergei

Thanks Sergei

Most of the genes which are overlapping are 'core' genes, which are highly conserved - I suppose in this case there won't be any bias in the analysis, as there isn't much variability.

I'm still analysing the sequences, but it looks to me that in the 'non-core' genes, the overlapping regions of the genes are much more conserved than the non-overlapping regions. Makes sense I suppose. Even in these genes, there aren't many codons predicted as being under positive selection.
Interestingly, almost all of the predicted recombination sites within these genes are in non-overlapping regions - all of this seems to be pointing me (I think) towards the conclusion that CMV 'evolves' by recombination rather than by drift/selection.

Actually, a bias towards finding recombination breakpoints in less conserved regions suggests to me that there is simply more power to detect breakpoints in those regions..

Cheers,
- Art.