BrassicaDB Peptide Process

Download SPTrEMBL from the EBI's FTP site. This database is built weekly so this will be an up to date copy.

Extract the Brassica records from the flat files (brass_embl.pl):

% zcat sptrembl/sprot.dat.Z | ~/perl/brass_embl/brass_embl.pl
% mv brass_embl.dat Brassica_SwissPROT.dat
% zcat sptrembl/trembl.dat.Z | ~/perl/brass_embl/brass_embl.pl
% mv brass_embl.dat Brassica_TrEMBL_.dat
% zcat sptrembl/trembl_new.dat.Z | ~/perl/brass_embl/brass_embl.pl
% mv brass_embl.dat Brassica_TrEMBL_new.dat

	SwissPROT	TrEMBL	TrEMBL new	Total
Selected Brassica peptides	162	479	186	827
Brassica peptides	216	552	212	980
Total records	84,893	235,493	117,496	437,882

Parse the files produced using the SwissPROT parsing script to generate the ace files.

% ~/perl/seq2ace/sp2ace/sp2ace_m.pl *.dat
% mv sp_seq.ace SPTrEMBL.ace

Parse the files and load them into a scratch database. Compare the numbers of objects with the production database.

	BrassicaDB	Scratch database
Peptide	758	823
Protein	758 (1,176)	823

The peptides from TrEMBL new could be a problem since their accessions will change when they are added to TrEMBL. This is fine if the proteins are going to be re-aquired each time, but not so great if incremental builds are prefered.

For now we will go with a re-aquisition stragegy, but this may change later.

So strip the peptide data out of the current database, and load up the new data.

Also generate a BLAST database of SPTrEMBL to use for analysis until next rebuild.

% zcat *.dat.Z | /usr/local/blast2/sp2fasta -g - > SPTrEMBL
% /usr/local/blast2/setdb -t SPTrEMBL SPTrEMBL