Interpreting results, 2D-sfs

[Daniel Portik]

Hi Ryan,
I've been testing a number of demographic scenarios for two populations of a particular taxon, and the best result appears to be a model involving a split with isolation followed by secondary contact and heterogeneous migration rates.  When inspecting the residuals, I see there is an interesting (and potential red flag) pattern that I haven't seen in the other taxa I've used previously. A little about this data set: there is a clear hybrid zone present, and after some discussion the leader of the project decided to remove the admixed individuals to minimize the confounding effects of contemporary hybridization. I'm not sure if this is having an unexpected effect on our inference and was hoping to get your opinion on the plots of our spectra. I can provide all the data and code if necessary.  Thanks!

Best,
Dan

[Gutenkunst, Ryan N - (rgutenk)]

Hi Dan,

That residual plot is definitely concerning. There is clearly something happening in the data to give you many mutations at around 50% frequency in both populations, and the model cannot account for that. Those might be coming from recent hybrids you haven’t eliminated. Or possibly it has to do with the data itself, perhaps paralogs in the genome that are being called as heterozygotes.

Sorry I don’t have an obvious explanation!

Best,

Ryan

--
You received this message because you are subscribed to the Google Groups "dadi-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dadi-user+...@googlegroups.com.
To post to this group, send email to dadi...@googlegroups.com.
Visit this group at https://groups.google.com/group/dadi-user.
For more options, visit https://groups.google.com/d/optout.
<atrox_West_vs_East_secasymmig.pdf>

--

Ryan Gutenkunst
Assistant Professor of Molecular and Cellular Biology, University of Arizona
phone: (520) 626-0569, office: LSS 325, web: http://gutengroup.mcb.arizona.edu

Latest papers: 
“Selection on network dynamics drives differential rates of protein domain evolution”
PLoS Genetics; http://dx.doi.org/10.1371/journal.pgen.1006132
"Triallelic population genomics for inferring correlated fitness effects of same site nonsynonymous mutations"
Genetics; http://dx.doi.org/10.1534/genetics.115.184812
"Whole genome sequence analyses of Western Central African Pygmy hunter-gatherers reveal a complex demographic history and identify candidate genes under positive natural selection"
Genome Research; http://dx.doi.org/10.1101/gr.192971.115

[Daniel Portik]

Hi Ryan,
I appreciate the fast response to my message. I wanted to make sure I wasn't missing something obvious in the interpretation. You've given me a couple clues about things to look for, and I'll go back to the data set with these ideas in mind.

Best,
Dan