Pindel error - malformed record detected

Question

[rooksie]

Hello,

I have sam2pindel input files that are giving me a malformed record detected error.

My command is:

pindel -f ${refdir}/${refname} -p ${filename}_recal_output4Pindel.txt \
 -c ALL -o ${filename}_recal -T 10 --report_long_insertions true \
 --report_breakpoints true --report_close_mapped_reads true

The output is:

Pindel
Initializing parameters...
Pindel version 0.2.5b5, 20150726.
Loading reference genome ...
Loading reference genome done.
Initializing parameters done.
SearchRegion::SearchRegion
Processing region: NC_006088.4  1   196202544
Chromosome Size: 196202544
NumBoxes: 60007 BoxSize: 6546

Looking at chromosome NC_006088.4 bases 1 to 5000001 of the bed region: chromosome NC_006088.4:1-196202544 
Scanning and processing reads anchored in NC_006088.4
Something wrong with the read name: 
malformed record detected in Sample32_recal_output4Pindel.txt

The pindel input file was generated using sam2pindel because the bam files were from bowtie2 alignment and then GATK was used to realign and recalibrate them.

Here is that command:

samtools view ../5_SNP_Indel/${filename}_recal_reads.bam \
   | ./sam2pindel - ${filename}_recal_Output4Pindel.txt ${length} ${filename} 0

I previously ran pindel using input files generated from bam files immediately after bowtie2 (before processing with GATK), and these files did not have this error. They ran successfully.

I suspect the issue may have arisen due to the realign and/or recalibrating of GATK.

Is there a way to check or correct whatever error(s) this is causing pindel?

Please let me know what additional information/data I can supply.

Thank you, Rooksie

Answer 1

[Kevin Blighe]

Try:

samtools view -h ../5_SNP_Indel/${filename}_recal_reads.bam \
   | ./sam2pindel - ${filename}_recal_Output4Pindel.txt ${length} ${filename} 0

The -h will ensure that the SAM header, which records contig information, is retained. May help.

Kevin