write Pindel error to text file

Question

[genya35]

Hello,

I'm running Pindel on small Ion Torrent (single-end reads) bam files. When the insert size specified in the config file is less than detected by the software, Pindel crashes with an error message. Could someone please suggest how to write that error message to a text file?

Here is the command that I use to run Pindel in my Perl script.

my $cmd1 = `$pindel -f $fastafile -i $configfile -c ALL -o $outputfolder`; print $cmd1;

Thanks

Answer 1

[Kevin Blighe]

Yes, I recently used pindel and it does crash when the insert size is not suitable. I presume that you're running it on a lot of samples.

Why not just redirect STDOUT and STDERR to a file, either appending to one 'master' output log or just one per sample? After that, you can just grep out the error message in question to infer which samples have failed.

./pindel [options] &> out.log

I am not aware of any functionality in pindel itself that does what you want.

Kevin

[genya35]

Sometimes Pindel overestimates the "insert size" which i assume is the read length for single-pair reads. Do you know? Do you know of any other good indel callers for Ion Torrent single-pair reads somatic(no matching normals)? Thanks

[Kevin Blighe]

Which indel size are you expecting? Ion Torrent is 75bp paired reads, right?

[genya35]

sorry, i mean single end and not pair reads. The amplicon size is 125, but error message sometimes goes up to 350. Doesn't make sense.

[Kevin Blighe]

Have you considered just running GATK? - it is pretty good for detecting indels, if not a little slow.

Like I said, I recently used pindel but it seems to find a lot of stuff that I feel represens false positive structural variants.

[genya35]

Pindel works well with KIT indels but misses some EGFR indels.

[Kevin Blighe]

Oh, you must be doing lung cancer (?). Today I actually used a program called FREEK, and it actually worked really well: http://boevalab.com/FREEC/

I think that it works with targeted data, assuming your data is targeted.

Unfortunately or fortunately, there are so many copy number programs out there. In terms of 'reputable' groups, including those where I work and also at Memorial Sloan Kettering Cancer Centre in New York, the programs people use are:

• pindel
• HMMcopy
• FREEK

That is, detecting copy number events in cancer NGS data.