Is BA3-SNPs sensitive to unequal sample sizes between populations?

[imogen-foote]

Hi,

I have a question about the background assumptions of BA3-SNPs. I couldn't find anything in the documentation to answer my question but apologies if it is there and I just missed it. Or perhaps it is so obvious that it doesn't need explaining.

Nonetheless my question is about whether or not the programme is sensitive to unequal sample sizes between populations? If I am comparing populations of unequal size should I be randomly subsampling the larger population to match the smaller population size or given the large number of SNPs is this unlikely to have too much effect.

Thanks in advance for any light you're able to shed on the issue.

[stevemussmann]

In general, unequal sampling can impact assignment tests so it's definitely an issue worth considering. I haven't personally run this program on datasets that have had drastically unequal sample group sizes, so I don't know how it would behave from experience. However, it might be worthwhile for you to conduct a test using your dataset to see if you get different results with equal vs. unequal sample sizes.

You might also consider reading Meirmans 2014 and Faubet et al. 2007 (if you haven't already done so). I don't recall them mentioning this issue, but it's been a while since I've read them myself, and I generally recommend them to people using any version of BayesAss.

https://onlinelibrary.wiley.com/doi/full/10.1111/1755-0998.12216
https://onlinelibrary.wiley.com/doi/10.1111/j.1365-294X.2007.03218.x

[imogen-foote]

Thanks very much - I will have a read of those papers and investigate this a little further.