Inferring divergence time from SNP data in SNAPP?

[Haley Arnold]

Hello! I'm very new to SNAPP so any guidance is appreciated. I followed along with the SNAPP tutorial but am not quite sure how to interpret the results. I am including my parameters and intermediate results which I would also appreciate any necessary guidance on.

I have SNP data from 3 diverging populations of a butterfly species, and would like to know very roughly when they began to diverge. I know approximately the age of the species and the mutation rate for a related species. I subset my data to 5 individuals per population. I converted my VCF to a nexus file, and loaded this into Beauti. 

Following the recommendations in the SNAPP tutorial, I set u and v mutation rates to 1, and did not sample. The coalescence rate was set to 10, and was sampled.

Following another paper's methods, I set the priors as pictured (no changes made to the u and v priors):

and then ran SNAPP.

Overall, I think my Tracer results were good, with the theta3 value low but everything else in the hundreds as recommended by the tutorial. 

Convegence of the trees in Densitree also looked good:

Finally, I used TreeAnnotator as described by the tutorial to produce a file for FigTree, shown here with the posteriors as branch labels and Node Bars set to display height_95%_HPD.

This looks great to me and supports the order of splitting populations as I expected.

However, I haven't yet found an explanation of the scale bar at the bottom (very new to constructing phylogenies) and don't know how I would go about interpreting divergence times. Wouldn't I have needed to incorporate the known mutation rate? Or the age of the species? Where would be the appropriate place to put that information in Beauti when generating the .xml file? I did read the documentation but was a bit confused as to how one should arrive at appropriate prior values.

Please let me know what steps I should take to learn this information, and any recommendations would be very helpful.

Thank you!

Best,

Haley

[michaelm]

Dear Haley,

Perhaps this tutorial might be useful: https://github.com/ForBioPhylogenomics/tutorials/tree/main/divergence_time_estimation_with_snp_data

With best wishes,

Michael