You are correct that ABBA-BABA requires four taxa, but it wouldn't work for your stated purpose anyway. ABBA-BABA is a whole genome test that can answer the question is there is evidence of introgression in your data. It does not, however, tell you which SNPs are introgressed. That is a much more difficult problem. f3 test is a 3-taxon test that can test if there has been introgression, but again, like ABBA-BABA it won't tell you which SNPs.

The problem is that alleles shared between two species could be due to lineage sorting (shared variation inherited from a common ancestor) or admixture.

You could look at the fd statistic in Simon Martin et al. 2014, which can be estimated in windows, but even this has problems. Another approach would be to infer introgression tracts using chromosome painting or PCAadmix, but these all come loaded with assumptions and often require phased data.

In short, your question is a challenging one and there are not easy answers.

Although ABBA-BABA requires 4-taxa, one of them has to be the outgroup. So if you have 3 ingroup species and an outgroup, then you can use it. Other than that, you can infer ancestry coefficients using STRUCTURE or other similar programs like ADMIXTURE or sNMF. Those can show admixture but doesn't provide direct estimates of gene flow. Also look at Treemix and fastsimcoal2. Here's one of my papers that estimates gene flow using RadSeq data:

Article Species delimitation with gene flow: A methodological compar...

Hope this helps. All the best!

An alternative for ABBA-BABA test could be f3 statistics (Peter, 2016) which allows to determine the level of "treeness" in your 3 species systems. Also ABC simulations like DIYABC could be useful to contrast demographic scenarios of asymetrical geneflow vs pure divergence among species. Finally, if you have a combination of "pure" and admixed populations within your systems you could check on INTROGRESS or BGC from Gromerpt, both may estimate levels of introgresion in a genomic cline-based approach

Thanks for all the answers! Sean Stankowski The 3 species I chose are in the same genus, having similar leaf morphology, and sometimes distributed sympatrically (but with slightly different niches). Species A normally distributed at middle elevation and the other two distributed at lower elevation. sometimes Species A could be found at low elevation, living sympatrically with species B and C. I'd like to test if adaptive introgression happened among these 3 species. So I have to know which SNPs or genes are under selection and introgressed between A and B or A and C. I used LFMM and other methods to find out the SNPs under selection, but I need to know if the selected SNPs are introgressed.

Jonathan Flowers Chan Kin Onn Marco Suarez-Atilano thanks for your recommendation, I'll read the articles or methods you listed.