Is dbSNP’s "ancestral allele" the same as HapMap's "reference allele"?

The NCBI "Reference allele" for a given SNP refers to the nucleotide base on the NCBI reference assembly at the SNP’s position. We think that this is what HapMap's "reference allele" refers to as well. We can confirm this if you will send us a few examples of where HapMap uses the term "reference allele".

The dbSNP refSNP report page does not directly use the term "reference allele". In the “Integrated Maps” section of the refSNP cluster report, there is a column called "Contig Allele" since "Contig Allele" is the allele on the contig at the SNP’s position. The "reference allele" is the "contig allele" value when the assembly type is "reference" in the map table. (Please Note: a reference contig's name starts with "NT_" and a Celera contig's name starts with "NW_".) The contig alleles on different assemblies are often the same, but not always. Rs3000 is an example of a SNP that has different contig alleles.

The ancestral alleles provided in a human refSNP cluster report is the SNP allele as found in the chimpanzee; I should mention here that not all SNPs in dbSNP have ancestral allele data. This data was provided by Dr. Jim Mullikin at NHGRI. There is more information about ancestral alleles in the ancestral allele subsection of the Data Content section of this archive.(01/04/08)

How do I find ancestral allele information in dbSNP?

You can get these data from two the tab-delimited table dump files on the dbSNP FTP site: SNPAllele.bcp and Allele.bcp. When the SNPAllele.ancestral_flag is 1, the allele_id is an ancestral allele. To get the actual allele, please use the allele table (Allele.allele_id and Allele.allele). The DDL for shared tables is located in the shared_schema subdirectory, and the organism table DDL is located in the organism sub-directory (for this example I’m using human as the example organism).

You’ll find that the rsID in “SNPAncestralAllele”, is stored as an integer in the first column of the table. Here’s an example to help you — say you want to find the ancestral allele for rs3:

1.

Look for the number “3” in first column of the SNPAncestralAllele table and you’ll find that the corresponding ancestral allele ID (located in the second column) is 7.

2.

Now, look in the Allele table for the allele letter that corresponds to
Ancestral allele ID = 7. In the allele table, the allele ID number is again located in the first column:

7 C 2003-02-22 01:11:00.0 2 2003-10-06 17:42:00.0

So you can see from this example, that the ancestral allele for rs3 is “C”. (03/08:08/08)

How do I download a flat file that contains the ancestral state of each SNP(when this ancestral state is known)?

You can the get ancestral state of each SNP (when this ancestral state is known) by using the following two tables from the FTP dump files:

1.

Go to the organism​_shared_data file located in the dbSNP FTP site, and download the Allele.bcp.gz file.

2.

Go to your organism’s (human, in this case) “organism data” file located in the dbSNP ftp site and download the “SNPAncestralAllele.bcp.gz” file.

3.

You can get the column definitions for the SNPAncestralAllele and Allele.bcp.gz tables from the dbSNP main table.sql.gz, which is located in the shared_schema file of the dbSNP ftp site. (11/16/07)

Determining Ancestral Alleles

Annotating Ancestral Alleles

Interpreting Ancestral Allele Field in refSNP Report