In article <321C9D26.537 at bbsrc.ac.uk>, Aengus Stewart <aengus.stewart at bbsrc.ac.uk> writes:
>> I was under the impression that accession numbers were unique across both protein and DNA
> but today completely at random I selected A25909 and pulled entries from embl and PIR.
>> I take it that acc nums are only unique within one type of sequence?
They are unique between the nucleotide databases (EMBL/GenBank/DDBJ) and
SWISS-PROT. 'A' accession numbers are assigned to sequences taken from patent
applications at EMBL.
All single letter accession number prefixes are used by the nucleotide
databases, except O, P and Q. P and Q are used by SWISS-PROT, and when they
are all used up SWISS-PROT will use accession numbers which do not conflict
with nucleotide ones.
EMBL - EBI