How to obtain ancestral alleles based on Hg19 genome coordinates?

Question

[devenvyas]

I'm trying to obtain ancestral alleles for almost 600,000 (i.e., Human-Chimp common ancestor) SNPs. A lot of them have dbSNP rs #s, but a lot of them don't.

I am particularly lost in how to obtain these data and am wondering if someone can point me in the right direction.

[devenvyas]

But how would I extract that data from the 1000 Genomes project?

That thread confused me a bit, so they are saying the ancestral alleles from 1000genomes are representative of the human-chimp common ancestor from Ensembl?

Also, I stumbled across ftp://ftp.ensembl.org/pub/release-82/fasta/ancestral_alleles/homo_sapiens.tar.gz, but I was unclear whether this is equivalent to the Human-Chimp ancestor. I was wondering if someone could help on this matter. Thanks!

[Jorge Amigo]

historically, primates' alleles were used as ancestral alleles for evolutionary reasons. dbSNP used the chimp's, and Ensembl used the macaque's when I first started working with these huge resources, a few years ago. right now, 1000g uses a combination of primates which allows them to provide a deeper (and, in my opinion, a more appropriate) ancestral description:

http://www.1000genomes.org/faq/where-does-ancestral-allele-information-your-variants-come

by the way, the link you provide is a fasta file. if you would like to get a particular base from there I would suggest using samtools faidx, although querying 1000g variants directly may be all you may need.

Answer 1

[Sishuo Wang]

The following link might be helpful I think.

1000 Genomes Snps' Ancestral Allele

[Sishuo Wang]

Also, the following paper uses a ML method to infer ancestral alleles which might be more complicated than the way mentioned in the link above.

http://genome.cshlp.org/content/22/12/2455.long