Convertig blastplus to blast output

[sammy07]

Hi all,

here is the next blast problem (and I'm actually not sure there is a solution for this). I need to perform a blast against nr with some species/gis excluded. Formatdb offers the possibility to create a database subset based on a positive gi list; however, if this list is huge, the blast takes very long.
The blast+ application offers both a positive and a negative gi list argument and performs much more efficient.
However, I use some programs which expect the old plain-text blastall output as input and fail to parse the blast+ output (which differs in some details from blastall output).

So, this is my question: does somebody know of a possibility to convert the blast+ output to the old blast output format?

Thanks in advance and best regards!

[maubp]

There is a reason that for years (*) the NCBI (and projects like BioPerl and Biopython) have been telling people not to try to parse the plain text BLAST output - the NCBI have a long history of tweaking the text presentation and thus breaking parsers.

Can you not use the tabular or XML output formats?

(*) I think it must be years by now...

[sammy07]

Thanks for your reply. I think you are perfectly right. It's just that some of the programs I use seem to expect plain text BLAST output. I might try to tweak the BLAST+ output so the parser can read it...

However, I'm having trouble with BLAST+ right now.

blastx -db nr -frame_shift_penalty 15 -query myquery.fa

throws an exception:

Error: NCBI C++ Exception:
"/am/ncbiapdata/release/blast/src/2.2.25/Linux64-Suse-icc/c++/ICC1010-ReleaseMT64--Linux64-Suse-icc/../src/objects/seq/../seqalign/Seq_align.cpp", line 670: Error: CreateDensegFromStdseg(): Std-seg segment lengths not accurate!

The query is:

>myquery
GTTCTACATAAGCAACAGGTATATGCCCAACGTTATTATTCTTCTATATAAAAGATAAGCAAATC

There is no exception if there is no frame-shift-penalty argument or if the sequence is different (e.g. the last 5 nucleotides removed). Someone knows what I'm doing wrong?

[maubp]

Thanks for your reply. I think you are perfectly right. It's just that some of the programs I use seem to expect plain text BLAST output. I might try to tweak the BLAST+ output so the parser can read it...

I remember having to tweak the Biopython parser to take BLAST+ plain text output.

However, I'm having trouble with BLAST+ right now.
...
Someone knows what I'm doing wrong?

I've not seen that before. I'm trying the example here (although it is likely our copies of the NR database won't be exactly the same).

Update - Same Error: CreateDensegFromStdseg(): Std-seg segment lengths not accurate! here from BLAST 2.2.25+ under both Linux 32 bit, Linux 64 bit and Mac OS X.

You should report this to the NCBI.

[maubp]

You should report this to the NCBI.

I've reported this by email.

Update: Wayne Matten at the NCBI replied this was a known bug, and the BLAST developers are still working on a solution.

[sammy07]

Great, thanks a lot!!

[sammy07]

BLASTX 2.2.26+
...
Error: CreateDensegFromStdseg(): Std-seg segment lengths not accurate!

Doesn't seem that it has changed with blast 2.2.26+. It's a pity...

[sammy07]

The problem was elegantly solved by the blast release 2.2.27+. The parameter -frame_shift_penalty no longer exists.