Codeml To Study Selection Pressure

Question

[GR]

Hi All,

Just few quick questions about Codeml.

Can we study lineage-specific selection only on sequences that are single copy orthologs (or in simple words just have 1-1 relations). What should I do in case of multigene families? Does codeml works well in case of multi gene families? Do I have to extract 1-1 relations from these families and then input to codeml or the whole family can go as input?

Thanks, RT

Answer 1

[DG]

The short answer is that, it really all depends on what you want to look at. There have been a series of papers out recently by Marc Robinson-Rechavi, Matthew Hahn, and a few others recently about selection and orthology/paralogy and the notion that orthologs should be more similar to one another functionally than paralogs. Ideally yes, I think you should extract the 1-to-1 ortholog relationships and look at selection pressures between those, but it really all depends a bit on what question you are trying to address.

[GR]

Thanks Dan. I know the best approach would be to identify 1-1 ortholog and then study the selection pressure.

In my case, I have 200 genes so it is difficult to identify 1-1 relation for all the genes. We just want to give a general statement whether our gene set is under selection or not. Do you have any useful reference for studying selection on entire gene families using codeml? and reference for codeml used for such a large dataset will also be a great help.

Thanks once again. Reetu

[SES]

Hi Dan, could you edit your post and include some links to those papers? I would find those papers very helpful. Thanks.

[DG]

This was the most recent paper "resolving" the ortholog conjecture: http://www.ncbi.nlm.nih.gov/pubmed/22615551

It references the Hahn and earlier papers so is probably the best starting place.

[SES]

Excellent! Thanks for the response.