Convert Sam file with incorrect fields to Fastq using HTSeq

[AnneBiton]

Hello,

I am working with BAM/SAM files generated by MapSplice on (human) RNA-seq paired-end reads from Illumina HiSeq. I want to convert them into fastq in order to realign them (I don't have access to the raw data).

To do this I am trying to use HTSeq-0.5.4p3 with python2.7 using
for a in HTSeq.bundle_multiple_alignments(samfile):
a[0].read.write_to_fastq_file(myfqfile)

The problem is that these Sam files have incorrect Sam fields and I get the error
ValueError: ("Malformed SAM line: MRNM == '*' although flag bit &0x0008 cleared").
The same error is raised with the picard tool SamToFastq.

I know this is a normal error since the Sam files are incorrect but I wanted to know if someone had a workaround for this problem?
I will eventually work on a script to reformat correctly all the Sam files but if someone has one I would be very happy to have it!

Thanks,
Anne

[maubp]

You could try 'samtools fixmate' or similar to deal with errors like this in the MRNM field, but it sounds like there could be a bug in MapSlice if it produces invalid SAM/BAM files.