Greetings GenBank Users,
The non-cumulative GenBank Update files for 11/26/99 and
11/27/99 contained records that lacked sequence version
numbers. The update files affected were:
nc1126.aso.Z
nc1126.flat.Z
nc1127.aso.Z
nc1127.flat.Z
Here's how the VERSION line appeared for an affected record
in nc1126.flat.Z :
LOCUS AF037745 183 bp DNA GSS 24-NOV-1999
DEFINITION AF037745 Homo sapiens SupT1 Homo sapiens genomic clone CONS81,
genomic survey sequence.
ACCESSION AF037745
NID g3722099
VERSION AF037745 GI:3722099
AF037745 on the VERSION line should be AF037745.1 . A total of
2475 records exhibited this problem: 2320 in nc1126, 155 in nc1127.
We patched all of the affected files (removing the records
involved) and installed new versions on the NCBI ftp site at
11:30pm on November 30th, 1999.
And prior to this, the 2475 affected records were re-distributed
via the nc1130 GenBank Update files (installed on our FTP site
at about 3:30am on November 30th).
If your local systems were unable to process the entire content
of nc1126 and nc1127 due to the missing version numbers, simply
re-transfer and re-process the patched files. If your local systems
were unable to process just those records that exhibited the version
number problem, then loading nc1130 will bring your system up
to date.
The problem once again stems from changes to the manner in
which GSS sequences are processed at NCBI. Unfortunately,
the fixes made earlier this month did not take into account
one particular source of GSS sequences and the method with
which they were moved from one database to another late
last week.
We are making changes to one of our "downstream" applications
which should be more effective in preventing future recurrences
of missing version numbers.
We regret this error and apologize for any inconvenience that
it may have caused.
Mark Cavanaugh
GenBank
NCBI/NLM/NIH
