How to run blast+

[mw55309]

Hi Dom

Sorry, I know this isn't a problem with blobtools per se....

My blast+ output looks like this:

tig00000001 N/A 54.3 tig00000001 N/A 53.9 tig00000001 N/A 53.1 tig00000001 N/A 53.5 tig00000001 N/A 53.9 tig00000001 N/A 47.8 tig00000001 N/A 27.7 tig00000001 N/A 53.1 tig00000001 N/A 52.4 tig00000001 N/A 53.1 tig00000001 N/A 52.4 tig00000001 N/A 53.1 tig00000001 N/A 51.6 tig00000001 N/A 53.1 tig00000001 N/A 52.8 tig00000001 N/A 52.8 tig00000001 N/A 53.5 tig00000001 N/A 53.1 tig00000001 N/A 52.8 tig00000001 N/A 52.8 tig00000001 N/A 52.4 tig00000001 N/A 52.4

Obviously N/A us not very useful.

Command ran was:

blastx -db nr_blastplus -query test.fa -outfmt '6 qseqid staxids bitscore' -num_threads 8

Building the database was:

zcat nr.gz | makeblastdb -in - -parse_seqids -dbtype prot -title nr_blastplus -out nr_blastplus

Have you seen this type of behaviour before?

Cheers Mick

[DRL]

Hi Mick, the reason for the N/A's is that either it can't find the TAXIDs of the sequences (because $BLASTDB is not pointing to NCBI's TaxDB) or there are no TAXIDs assigned to the sequences in the DB.

Is this NCBI's nr or a custom subset?

• If it's NCBI's nr you have to download NCBI's taxdb and set the $BLASTDB variable

  wget ftp://ftp.ncbi.nlm.nih.gov/blast/db/taxdb.tar.gz
  export BLASTDB=$BLASTDB:~/FOLDER/WHERE/TAXDB/LIVES

• If it's a custom BLASTDB, you have to generate a taxid_map file (SEQID\tTAXID\n) which you pass on to makeblastdb

  makeblastdb -in blastdb.fasta -dbtype nucl -taxid_map blastdb.id_to_taxid.txt -parse_seqids

I just had to do something similar where I wanted to construct a new BlastDB with ONLY subjects hit by a previous search against NCBI nt:

# extract subject_ids from blast result (column 5 is sseqid in my case)
cut -f5 megablast.out | sort | uniq > subject_ids.txt
# Get tax_id_map 
blastdbcmd -db nt -entry_batch subject_ids.txt -outfmt '%i %T' -out subject_sequences.tax_id_map
# Get sequences
blastdbcmd -db nt -entry_batch subject_ids.txt -out subject_sequences.fna
# make DB
makeblastdb -in subject_sequences.fna -dbtype nucl -taxid_map subject_sequences.tax_id_map
 -parse_seqids

Hope that helps.

cheers,

dom