BLAST Process

For BrassicaDB we perform a range of BLAST analyses:

BrassicaDB ESTs vs. SPTrEMBL
BrassicaDB ESTs vs. BrassicaDB protein
BrassicaDB cDNA probes vs. EMBL

For each set of sequences to be BLASTed we dump them out of the database in fasta format. These files are transfered to the cluster along with a copy of the BLAST databases (EMBL, SPTrEMBL, BrassicaDB protein).

The data sets are generated using the following queries:

BrassicaDB ESTs
```
find Sequence_EST *; follow DNA
```
BrassicaDB cDNA probes
```
find Sequence * Probe; follow DNA
```
BrassicaDB protein
```
find Peptide *
```

All the BLASTs are performed using WU-BLAST 2.0.

Currently we use a small Linux cluster to perform the BLASTs.

wu-blastn.pl
Used for all the normal BLASTNs.
wu-blastx.pl
Used for all the BLASTXes.

The BLASTs are performed in two runs: one for EMBL and SPTrEMBL and one for the BrassicaDB BLASTs.

A small script (rebuild.sh) is used to automate each run.

% ./rebuild.sh

The files produced are just a long list of BLAST reports. For the purposes of parsing these into ace files a simple script (lc2blasts.pl) is used to seperate them into individual BLAST reports (one to a file).

% ~/perl/seq2ace/blast2ace/lc2blasts.pl ../archive/Brassica_ESTs.blast SPTrEMBL

The BLAST parsing script is then run over the the BLAST reports and an ace file produced. This ace file is then read into the database.

% ~/perl/seq2ace/blast2ace/blast2ace.pl BLASTX_SPTREMBL SPTrEMBL "*" EST_SPTrEMBL_BLASTX-20000407.ace similar protein 1 "" 0

Note: remember to check the ace file is correct before reading it into the database.