dated species tree

[roberto feuda]

Hi Benoit,

I hope you are doing great! I have a question about the DTL model. I am running the DTL model using the per species rate. Still, the reconciled tree infers the transfers between species that didn't exist at the same time (they are biologically implausible). Do you have suggestions on how to correct this? 

Best wishes,

Roberto 

[Benoit Morel]

Dear Roberto,

Sorry for the late reply, this mail ended up in my spam folder.

Unfortunately, GeneRax cannot forbid "impossible" transfers (between non-contemporary species) because it does not account for the species tree branch lengths. The per-species rate is not expected to help further here.

Note that if those transfers go down (from an older to a more recent species), then the transfer is still plausible, but an extinct species might have been involved in the process. Some studies suggest that this might have happened a lot (I think this paper explores this idea: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3622898/). But if the transfers goes up, then it's very likely an inference error.

The only suggestion I can provide is to use a tool that forbids impossible transfers. For instance, ALE provide such models, assuming that you can trust your species tree branch lengths: https://github.com/ssolo/ALE

I hope this helps,

Benoit