Are branch-specific DTL rates allowed in GeneRax?

[Yiyan Yang]

Dear all, 

Thank you for providing such a useful and convenient tool. I am wondering if it i spossible to estimate DTL rates for different branches? For example, I want to assign 3 set sof DTL rates for 3 types of branches in the tree (and different types of branches are with different traits) and compare the DTL rates among them.

If this is possible , that would be wonderful!

Thanks,

Yiyan

[Benoit Morel]

Dear Yiyan,

Yes and no :-)

There is an option --per-species-rates which estimates a different set of DTL rates for each species tree branch. However I haven't used it for a long time and it's not officially supported. In particular, I don't think that the per-branch DTL rates are outputted anywhere (they are just used to estimate the likelihood score and to reconcile the gene trees).

I have a few questions, to understand how I can make this work for your use case:

- do you need GeneRax to estimate the per-species-branch DTL rates, or do you want to set them yourself?

- should I output the per-species-branch DTL rates in some file?

Also, note that there is per_species_event_counts.txt file that contains the number of events per species branch. (but I understand that this is not exactly what you are looking for)

Best,

Benoit

--
You received this message because you are subscribed to the Google Groups "GeneRax" group.
To unsubscribe from this group and stop receiving emails from it, send an email to generaxusers...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/generaxusers/06c1dde0-2780-4b38-8152-9a3f6753f3b8n%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Yiyan Yang]

Dear  Benoit,

Thanks so much for your rapid response.

- do you need GeneRax to estimate the per-species-branch DTL rates, or do you want to set them yourself?

I would love to let GeneRax to estimate the per-species-branch DTL rates. But for me, this sounds more like a "free-rate" model for every branch. I am wondering if I could constrain a group of branches with all of them having the same DTL rates but with another group having a different kind of rates? (For example, branches 1-10 have the same D=0.1, L=0.1, T=0.1, S=1 estimated by GeneRax and branches 11-20 also have the same rates but with D=0.5,L=0.5,T=0.1,S=1. And this is more like a "partially-free-rate" model which is helpful to test evolutionary hypotheses in comparative biology.) 

- should I output the per-species-branch DTL rates in some file?

It will be very helpful! If it is available, please let me know. Thank you!

I think a "free-rate" model is good enough and I also noticed a "global-rate" model is alreday supported in GeneRax. I may ask for too much for a "partially-free-rate" model and it may take some time and effort to add this feature/function. So I guess I will just focus on the per-species-branch DTL rates first but eagerly look forward to a "partially-free-rate" model.

Really appreciate your help and I am again amazed that this tool is well maintained by you and your group.

Best,

Yiyan

[Benoit Morel]

Thank you for your nice feedback :-)

I will try to output the per-branch rates and to check that the "free rates" works properly in the next few days. Let me know if you are installing via Bioconda, because I would have to update the package. Do not hesitate to ping me if you don't get any news from me after one week.

The "partially free rates" is a bit more complicated. I need to think about it before telling you if I can implement it.

Benoit

To view this discussion on the web visit https://groups.google.com/d/msgid/generaxusers/15492342-1e0b-4eb2-878e-4908bd4d52c6n%40googlegroups.com.

[Yiyan Yang]

Thank you. This is awesome! Yes, I installed GeneRax via bioconda. Many thanks again.

Have a great day!

Best,

Yiyan

[Benoit Morel]

Dear Yiyan,

GeneRax now outputs the per-species rates (when --per-species-rates is enable). See this part of the wiki: https://github.com/BenoitMorel/GeneRax/wiki/GeneRax#outputs

I am updating the bioconda package, but it might take one day to be published (it has to be manually approved by the bioconda team). I will let you know once it's available.

Don't hesitate to let me know if the estimated rates make (or do not make any) sense. The feature is not really new, but it hasn't been as well tested  as the rest of the program.

Best,

Benoit

[Yiyan Yang]

Dear  Benoit,

Thanks so much for your email. This is fantastic! Thank you again for your effort and help.

Thanks,

Yiyan