For benchmark purposes I’m trying to find an old version of the human reference genome (pre-GRCh37) and a matching version of dbSNP.

Unfortunately it appears as though older versions of dbSNP aren’t archived, or at least these archives are not publicly accessible.

In particular, the NCBI FTP archive is fairly sparse. The only pre-GRCh37 version it contains is ostensibly dbSNP build 124 from January 2005, for hg17 (NCBI build 35). Unfortunately the corresponding FTP folder is actually empty. dbSNP build 138 has been “removed to recover storage”, whatever that means, and I’m unclear which version of the human reference genome it refers to (its official page does not contain this rather crucial bit of information).

Is there another authoritative source of this data?

  • 1
    $\begingroup$ Perhaps ENSEMBL will have some of what you need? At least the older reference genomes appear to be there. ensembl.org/info/data/ftp/index.html $\endgroup$
    – Scot
    Commented Apr 9, 2019 at 5:57
  • $\begingroup$ @Scot Ensembl has variation data but it does not, as far as I know, host the dbSNP data (i.e. a dataset exactly corresponding to a given dbSNP build). Although I might be able to filter the corresponding entries from their variant database, and convert it to VCF. $\endgroup$ Commented Apr 9, 2019 at 8:28
  • $\begingroup$ Actually this won’t work: The NCBI36 variant data on Ensembl doesn’t contain dbSNP information directly, and it’s in a format that’s not trivial to convert to VCF. $\endgroup$ Commented Apr 9, 2019 at 8:58

1 Answer 1


The UCSC Genome Browser downloads page has old human genome assembly versions way back to the first human genome ever published (by UCSC in 2000). http://hgdownload.soe.ucsc.edu/downloads.html#human

We also have a variety of archived dbSNP data up to dbSNP141 on hg38, dbSNP 131 on hg37, dbSNP126 on hg18: http://hgdownload.soe.ucsc.edu/goldenPath/hg38/database/snp141.txt.gz http://hgdownload.soe.ucsc.edu/goldenPath/hg19/database/snp138.txt.gz http://hgdownload.soe.ucsc.edu/goldenPath/hg18/database/snp126.txt.gz

They're in a custom table format, which can be turned into VCF manually.

You can also get all of this data from Table Browser and add filters, etc.: http://genome.ucsc.edu/cgi-bin/hgTables


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.