Can blastp (blast+) print a line for seqs with no hits when using -outfmt 6?

[kmkocot]

Hi all,

I'm trying to write a script that processes blastp output when using the -outfmt 6 setting. I want blastp to print a line for sequences with no hits. It doesn't matter what the line says as long as the name of the query sequence is printed. I thought about making the e-value really high but some sequences will still hit nothing. Is this possible?

Example command:
blastp+ -db good -query input.fa -num_descriptions 1 -num_alignments 1 -outfmt 6 -num_threads 6 > intput_vs_good.txt

Thanks!
Kevin

[SES]

In order to determine which reads have no hits you simply create a file of sequence IDs with a BLAST hit (call this set i) and create another file with all your sequence IDs in the query file (call this set j). Next, sort the files (it is difficult to optimize the sort command without seeing your sequence IDs) and then compare the files to find the IDs in set j that are not in set i. That should help you design your script, but personally I would not write a script for this job. If you are familiar with the unix tool chain you can do all this in a couple of commands with cut, grep, sed, sort, and comm.

[maubp]

In order to determine which reads have no hits you simply ... and create another file with all your sequence IDs in the query file (call this set j)...

I find the input query FASTA file works very well for this if writing a simple script.

[SES]

I find the input query FASTA file works very well for this if writing a simple script.

That is correct. I was referring to the query Fasta file when I mentioned the "query file." Sorry if that was not clear. There is one more dependency that I had not mentioned with a command line approach. That is, you would need some program for indexing the query Fasta and extracting your IDs that did not appear in the BLAST report (e.g., cdbfasta, Vmatch, leaff, etc.).