Simulate locus tree under assumption of whole genome duplication

[Liming Cai]

Hi Diego,

I am interested in the the effect of genome duplication and subsequent gene loss/gene conversion on phylogenomic reconstructions. SimPhy is capable of simulating locus/gene genealogy under incomplete lineage sorting + gene duplication/loss + gene conversion. But is seems that the gene duplication and loss is modeled in a chronologically homogeneous way. My questions is that is there any way that I can incorporate a whole genome duplication framework in the locus tree simulation? In other words, for species underwent whole genome duplication at certain time point, all of its descents get duplicated immediately after the event (with some level of immediate gene loss). At the same time, I would also like to take incomplete lineage sorting, gene conversion, and rate heterogeneity into account. Can this be achieved by SimPhy? Or if there are other programs that I can use jointly with Simphy to achieve this goal? My ultimate goal is to simulate the sequence alignment and conduct species tree inference on that and then compare it to the true species tree. 

Best regards,

Liming

-- 

Liming Cai

Graduate Student, Davis Lab

Dept. of Organismic and Evolutionary Biology

Harvard University

22 Divinity Ave, Cambridge, MA 02138

[Diego M.]

Dear Liming,

I have just realized that this has been waiting here for more than a year. I am very sorry for that. 

The problem you are working on is very exciting, and SimPhy has in place all the machinery needed to simulate this scenario. However, as it is, this is impossible to do (either alone or in combination with other software). You would need to modify SimPhy's code (adding tags/nodes in the species tree that indicate genome duplication events, a fourth "genome tree" between the species and the locus tree, or species-branch dependent locus tree simulation parameters). I do not have the time right now to do it on my own, but I would be happy to help/collaborate with you (or anybody interested) in order to implement this in SimPhy (or a derived new software).

Best,

Diego M.