SAM(tools) and BLAST

[fungs]

Hi there,

I am currently trying to use different aligners for metagenomic short/medium sequences and SAM seems to be a good intermediate format for my analyses. For sensitivity reasons my current reference is based on Blast but I cannot convert any blast result file into SAM format.

The perl script contained in the samtools release 0.1.17 called blast2sam.pl does not really work and isn't very informative. Anybody has experience with Blast and samtools? I tried different blast output formats but found none of them working.

-------------------------------------------------------------------------

> blast2sam.pl test.blastn
Use of uninitialized value $qend in subtraction (-) at /local/programs/samtools/blast2sam.pl line 58, <> line 1823.
Use of uninitialized value $qlen in subtraction (-) at /local/programs/samtools/blast2sam.pl line 58, <> line 1823.
Use of uninitialized value in substr at /local/programs/samtools/blast2sam.pl line 63, <> line 1823.
Use of uninitialized value in concatenation (.) or string at /local/programs/samtools/blast2sam.pl line 63, <> line 1823.
Use of uninitialized value in bitwise and (&) at /local/programs/samtools/blast2sam.pl line 65, <> line 1823.
Use of uninitialized value $sam in join or string at /local/programs/samtools/blast2sam.pl line 72, <> line 1823.
Use of uninitialized value $sam in join or string at /local/programs/samtools/blast2sam.pl line 72, <> line 1823.
Use of uninitialized value $sam in join or string at /local/programs/samtools/blast2sam.pl line 72, <> line 1823.
255 M * 0 0 * *

[lh3]

BLAST support will be dropped unless someone want to maintain it. I realize that it would be better to have fewer functionality to avoid letting others blame me for having too many bugs. I just thought this script may be useful to someone occasionally, but it is now causing more troubles than good. Sorry.

[fungs]

more verbose

I was aware that the script is not mature but for understanding its functionality I need some more information on its usage, e. g. what the BLAST output format should be for the program input. I was hoping someone could give me a hint into the right direction before starting to read through the perl code.

Since I made my decision for the SAM format I might as well end up writing my own converter from blast to SAM.

How about putting the script into the svn but not releasing it in the final tar balls?

[Heikki]

blast2sam.pl script works for me on default verbose textual output from blastall 2.2.19 .

SAM format is well documented and easy to produce. BioPerl has a solid parser for blast output. How about writing an output module for SAM format? Bio::AlignIO::sam would be great to have!

[lh3]

@fungs

blast2sam.pl works with the default output. It fails probably because some regex matching fails. As you have that blast causing the problem, it may be easier for you to debug it (around line 1823). If you can fix it, please let me know. Thank you.

[cjfields]

I think this has been implemented via Bio::Assembly::IO::sam.

blast2sam.pl script works for me on default verbose textual output from blastall 2.2.19 .

SAM format is well documented and easy to produce. BioPerl has a solid parser for blast output. How about writing an output module for SAM format? Bio::AlignIO::sam would be great to have!

[fungs]

Native support

Just to complete this: NCBI Blast+ now supports SAM directly.