Interpreting Residuals

[Sarah Babaei]

Hi Ryan,

Firstly, thank you so much for all your helpful responses so far! 

I have started using dadi-cli for demographic testing of my 2 population models. I have tested about 14 models for each (don't know too much about the system), and plotted the fit of the best model. The parameters seem to make sense and indicate expansions in all population pairs. I am now seeing some odd patterns in the residuals and was hoping to get your thoughts on them.

The asym_mig_size seems to be underpredicting singletons, does this point to another population expansion? 

Then the sec_contact_asym_mig_size model shows underprediction of shared alleles, which I'm confused about. These populations generally have small Nes, which I understand can be a problem.

Thank you,

Sarah

[Sarah Babaei]

Apologies for the double email, I thought I should include the .bestfits files for each. 

[Ryan Gutenkunst]

Hello Sarah,

Neither of these residual plots shows strong underprediction of singletons, son I’m not sure what you’re looking at there.

The sec_contact model on these data is underpredicting shared low-frequency alleles, although not too much. Maybe there’s a recent intensification of gene flow? I’m not sure these data would have power to confidently show that, but it might explain a good fit to the moderate frequency shared alleles but a poor fit to the lowest-frequency.

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 view this discussion visit https://groups.google.com/d/msgid/dadi-user/22e472df-81a9-44b1-a7bb-4fbbdb0bc7dan%40googlegroups.com.
<sec_contact_asym_mig_size.pdf><asym_mig_size_DNVEN.pdf>