BLAST

This is the Title of the Book, eMatter Edition

Copyright © 2012 O’Reilly & Associates, Inc. All rights reserved.

Sequence Databases

|

203

a mixture of letters, numbers, and other symbols, but not spaces.) This tag uniquely

identifies the sequence in a database. There isn’t necessarily a one-to-one correspon-

dence between sequences and tags because sequences are sometimes known by mul-

tiple unique names. The DDBJ/GenBank ACCESSION (or AC in EMBL) is the

primary name for a sequence record. Another unique name is the LOCUS (or ID in

EMBL). The locus is supposed to be a “short mnemonic name for the entry, chosen

to suggest the sequence’s definition.” For example, “HSMG01” is the locus name for

the database entry containing Homo sapiens myoglobin exon 1. Over time, like the

names of celestial objects, locus names have become less descriptive and are often

just duplicates of the accession numbers.

Sequence records can also change over time. This often happens when the record is

edited to correct a sequence error. The accession number and locus don’t change,

but the version number is increased (VERSION in DDBJ/GenBank and SV in

EBML). In this way, an ACCESSION.VERSION points to a particular record at a

particular time. It’s a good idea to always refer to sequences in this way and not by

ACCESSION alone or by LOCUS or ID.

DDBJ/GenBank records include an additional token called the GI number, which is a ...