ABBA BABA test for two population

Question

[reza]

I generated a VCF file including 20 samples belonging to two species (A and B), variant calling was done using GATK best practice pipeline (only autosomes chromosomes). Now I want to detect the introgressed regions from A sp. to B sp. I searched for a way to do that and find more articles used the ABBA BABA test and fd statistics. When I read some articles and some manuals for this test, I found they use 4 populations (P1, P2, P3, and O as outgroup).

Now my question is, how I can handle this analysis using two populations?

Do you have a suggestion for another valid method of identifying introgressed regions?

[4galaxy77]

ABBA BABA needs 4 populations by definition, so you won't be able to do it using only 2. I couldn't say if you could use another method, but hopefully someone else will.

Answer 1

[Vic]

You may have already thought of it but, is there any supplementary data you can download from previous studies? There is also the possibility to split a sample up? This is a good tutorial for ABBA BABA. But if you only have the two sample and no outgroup, its going to be difficult. There is a three way test discussed here.

[reza]

Thanks, Vic

What features should the outgroup have? For example, to identify the gene flow between cattle and yak, which species can be a candidate for the outgroup?

If I use the D3 for our work by adding another breed to my work, it tells me if there is an introgression or not, what should I do to identify the location of introgressed regions on the genome?

[Vic]

I don't know about what outgroup you should use, I don't work on cattle or yak, have a look at some phylogenetic trees, you need a good knowledge of your study species to decide what to do.

The D statistic will let you know if there is introgression, the the f (fd) statistics will allow you to investigate the overall proportion of admixture and you can also perform that on a sliding window along the genome to locate loci. Have a look at Dsuite .

Good luck!