This is a test version of Biostars. For the public version, visit https://www.biostars.org.
How to polarize ancestral versus derived alleles?

Hello biostars,

I am trying to get into popgen analysis in angsd and currently working on some summary statistics. This made me to think a little bit more about allele polarization for D-stat, f3, f4 stats, SFS and other analyses.

For D-statistic estimate, ANGSD is asking for ancestral fasta file. However, I am not sure what kind of fasta it should be. If all my BAMs are aligned, let's say, to hg19, but an outgroup is chimp, should I provide a reference PanTro genome? In this case, coordinates are different: BAMs are aligned to hg19.

Or, should I convert PanTro bam file aligned to hg19 into some kind of consensus fasta? Or, finally, I can realign all bams on chimp genome, and then use these realigned bams together with PanTro for the analysis. What is the best way to do that?

I guess it would be better to use a real outgroup to polarize alleles (especially when doing SFS), but some papers (as this one) use non-outgroup reference and do that using folded SFS with no problems.

In general, is there an optimal strategy for this kind of popgen decision making?

next-gen snp
• 6,233 views • link
updated by jena • jena • written by Alice Alice

Apologies that no-one else has responded. It is a very specific type of analysis that you are aiming to do, but very interesting I must admit.

From what I can see, ANGSD could accept a BAM aligned to hg19 and another aligned to the Pan troglodytes, however, this may not necessarily be the correct way to run the program.

I noticed this recent study, which appeared to run ANGSD separately on 3 different species: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4788117/

• 0 views • link
written by Kevin Blighe Kevin Blighe

Thanks! Yeah, it is not an easy question. I ended up aligning chimp on hg19. Other part of my question is very theoretical, I looked through the literature to see what people do - and they do whatever data allows. Some for examples do not have a sequenced outgroup so they just use a reference.

• 0 views • link
written by Alice Alice

Hi Alice, I am doing unfolded SFS, I didn't know how to use a real outgroup to polarize alleles, can you give me some suggestions?

• 0 views • link
written by qing • qing

I have a few crude scripts in my repo that can polarize VCF based on outgroup and calculate unfolded SFS (I call it "derived SFS" as that name is clearer to me). I'm currently adapting the code to use on another taxon and I see there are several assumptions about the format that need some work before it can be used. (For example, my VCF data for chimp were phased and GT-only, so the script currently does not handle non-phased format, or will waste time "fixing" (corrupting, really) other info like DP. The necessary changes are not big though, so if I get some better versions I will update the scripts. Or you can fix them yourself, it's rather simple AWK code.)

• 0 views • link
written by jena • jena

1 answer

I know this is quite late, but here is my experience working on this a few years ago:

Several primate genomes are directly available in the hg19 coordinates. I used several of them in the past.

For example, the SGDP-lite dataset contains a fasta file of chimp in hg19 coordinates (specifically in this 126GB tar archive). I think it used PanTro2 assembly in the past, but Shop Mallick (from Harvard) was updating it to PanTro6 last time I was in contact with him a few years back. You can probably just email him to get it directly (they have it on a FTP server, I just don't know if I can share the link here). I also got other primate genomes in hg19 coordinates from him, but I suspect they are publicly posted somewhere too.

I also see that Alan Rogers has scripts to download chimp (PanTro5 in hg19 coordinates) and gorilla from UCSC in his github repo.

Hope it helps anyone still looking for this.

• 0 views • link
written by jena • jena

Log in to answer this question.

More posts like this