I want a list of all variants, i.e. sites which are known to vary between human to human. For example, it should ideally cover all sites in here, but without samples.

I don't want a giant reference panel with many samples, nor a .fa which has no major/minor allele information.

I just want a file with the following information at each locus:

  • Location (chromosome position and chromosome)
  • Bonus (not necessary): rsID
  • Major allele (i.e., A, C, T, or G) at that position
  • Minor allele at that position

I'm not picky about how major/minor alleles are defined. Usually they are defined as more common in the general population vs. less common (loosely). Conforming to b37 or b38 would be ideal.

I'm looking for one such file that doesn't have samples in it, but does contain all possible positions (not all VCF files have this)

Where can I download something like this? I like working with VCFs but any format is good.


2 Answers 2


Sounds like you're looking for the dbSNp files. You haven't specified a species, but you can download the data for human from the dbSNP FTP site. Specifically, you want:

wget https://ftp.ncbi.nlm.nih.gov/snp/organisms/human_9606/VCF/00-All.vcf.gz

This file has lines like:

1   10848   rs1271898990    C   G   .   .   RS=1271898990;RSPOS=10848;dbSNPBuildID=151;SSR=0;SAO=0;VP=0x050000020005000002000100;GENEINFO=DDX11L1:100287102;WGT=1;VC=SNV;R5;ASP;TOPMED=0.99998407237512742,0.00001592762487257
1   10851   rs1340958514    G   A   .   .   RS=1340958514;RSPOS=10851;dbSNPBuildID=151;SSR=0;SAO=0;VP=0x050000020005000002000100;GENEINFO=DDX11L1:100287102;WGT=1;VC=SNV;R5;ASP;TOPMED=0.99996018093781855,0.00003981906218144
1   10854   rs1222740568    G   A,C .   .   RS=1222740568;RSPOS=10854;dbSNPBuildID=151;SSR=0;SAO=0;VP=0x050000020005000002000100;GENEINFO=DDX11L1:100287102;WGT=1;VC=SNV;R5;ASP;TOPMED=0.99967348369011213,0.00019113149847094,0.00013538481141692
1   10857   rs1487378566    C   G   .   .   RS=1487378566;RSPOS=10857;dbSNPBuildID=151;SSR=0;SAO=0;VP=0x050000020005000002000100;GENEINFO=DDX11L1:100287102;WGT=1;VC=SNV;R5;ASP;TOPMED=0.99933900356778797,0.00066099643221202

However, it sounds like you might be underestimating the complexity here. There isn't just one minor allele, many sites have multiple minor alleles. The example above shows one such multiallelic site (you can split these into separate lines using bcftools). But you also have cases like:

1   10786   rs1464509424    CGGCGCA C   .   .   RS=1464509424;RSPOS=10787;dbSNPBuildID=151;SSR=0;SAO=0;VP=0x050000020005000002000200;GENEINFO=DDX11L1:100287102;WGT=1;VC=DIV;R5;ASP;TOPMED=0.99991239806320081,0.00008760193679918
1   10786   rs1258377526    CGGCGCAGGCGCAGACACATGCTAGCGCGTCGGGGTGGAGGCGT    C   .   .   RS=1258377526;RSPOS=10787;dbSNPBuildID=151;SSR=0;SAO=0;VP=0x050000020005000002000200;GENEINFO=DDX11L1:100287102;WGT=1;VC=DIV;R5;ASP

Don't expect just one major and one minor allele per location.

  • $\begingroup$ This is probably it, thanks. I looked on the dbSNP website but couldn't find the exact right file. I have a good method in place for handing multiallelic sites $\endgroup$
    – BigMistake
    Oct 10, 2023 at 18:22

I agree with @terdon that perhaps dbSNP is what you may want to look at. The dbSNP155 RefSNP VCF file in GRCh37 version may help. It can be downloaded from https://ftp.ncbi.nlm.nih.gov/snp/latest_release/vcf/GCF_000001405.25.gz.

You may also fine the 1000G genome website helpful. https://www.internationalgenome.org/data-portal/data-collection contains much of its release both in .fa format and in VCF format. I understand that a large number of samples included increase the file size tremendously, but it is necessary as you're not just looking at what the major and minor allele is, but also their frequency, which may vary greatly between populations. It's important to check which population we're seeing when using their genomic information.

Hope my answer helps. Regards.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.