Blastp not matching to reference more than 1000 entries.

Question

[akshay_ware]

We are trying to create a database using local-blast makeblastdb command from fasta file containing 190000 lines (95000 fasta input). when we try to run blastp using this reference we got no match found for sequence even though the sequence is present in the reference. To troubleshoot we took parts of reference and its matching to the reference size upto 1000. how to increase blastp search reference size?

[Michael]

What are the exact commands used for makeblastdb and blastp? If a sequence is present exactly, it should definitely be found with an e-value close to 0, unless it is too short, the db too big, the e-value cutoff is too conservative or something else went wrong (there are a quite a few different things to mess up). In this case, I suspect that you are using a very short query and therefore the search yields significant hits only with a small DB.

[akshay_ware]

Yes, we are using very short query the command for creation of the database is :

makeblastdb -in ref.fasta -parse_seqids -titile "test" -dbtype prot

and blastp command is:

blastp -query query.fasta -db ref.fasta -word_size 2

The query length is five amino acids

[Michael]

-titile "test" should be -title "test" Just in case you might have overlooked the error message, you didn't get your database rebuilt.

Answer 1

[Michael]

ok, how long is the query effectively? And are you getting no hits at all or is the desired hit just not among them?

In the first case try setting -evalue 10000 or so, in second case change -num_descriptions and -num_alignments to 10000 or so.

[akshay_ware]

Hi, By adding the -evalue 10000 the query returned with the right match. Thank You!