Hi,

It has been done on Heliconius butterflies I think, but was criticized in the following paper by Cruickshank and Hahn.

Article Reanalysis suggests that genomic islands of speciation are d...

An issue with IMa is that it overestimates gene flow when divergence time (in Ne units) is recent. Also, it is quite time-consuming and you can probably use more recent methods to extract more information from your dataset. Two widely used methods include dadi and fastsimcoal, which have the advantage of allowing comparison between arbitrarily complex models. This way you can test (for example) secondary contacts v. constant gene flow, and there are dadi scripts that can even estimate the proportion of your genome affected by background selection or that constitute barriers to gene flow. You can check this paper:

Article Adaptive evolution and segregating load contribute to the ge...

If you still want to extract sequences and use IM, I would suggest you use the perl script vcf-consensus in vcftools to generate a consensus sequence for each individual haplotype, providing your VCF is already phased. https://github.com/vcftools/vcftools/tree/master/src/perl

Note that the script would have to be run for each individual haplotype first, but you can easily include it in a loop.

Hope that helps a bit

Best,

Yann

Hi Yann,

Thank you so much for your feedback and advice, it is really helpful. I realized that IMa2 is not the best software indeed for such analysis so I would give a try with fastsimcoal2.

By the way thanks for your awesome and helpful review on population genetics software, I recommend it to all my colleagues who don't know where to start!

Best

Maeva

Yes, fastsimcoal is quicker to understand and use (at least for me). And you can couple it with Arlequin to get directly AFS files in the right format.

Thank you for the review! I am working on a new draft, I will resubmit it soon. There are improvements in machine learning that I did not consider at first but that might become interesting to apply even on "non-model" species.

Good luck with you research!