samtools manipulation of bowtie output with suppressed fields

Question

[Sjannie Lefevre]

Hi

I am trying to create alignments from an riboseq experiment using bowtie1 (Bowtie/1.2.2-foss-2018b) on a HPC Linux cluster. Goal is downstream analysis using the riboseqr package.

Riboseqr expects sorted bam files created with '--suppress 1,6,7,8', or txt files (i.e. sam).

However, when I include this in my bowtie command, samtools (SAMtools/1.9-foss-2018b) gives me a 'truncated file' error when trying to sort or view the file.

I have succesfully created bowtie/samtools pipes before on the same data. I cannot find any information about whether samtools cannot read bam files that do not have all the standard fields - is it really so, and how does one get around it?

I have used ribosegr on ribogalaxy without problems, but there the file is converted to sam anyways. As riboseqr can also take sam files as input, one solution is to output sam files from bowite without --supress, then sort them, and then use sam as input for riboseqr (suppressing the uneeded columns directly, rather than in bowtie). But do to the larger size of sam files, it would obviously be more convenient to use the bam files.

Does anyone have experience with similar issues?

Cheers Sjannie

Answer 1

[Sjannie Lefevre]

Ah, I found the information about essential fields in SAM/BAM files (http://samtools.github.io/hts-specs/SAMv1.pdf), and it seems that 11 are mandatory, so that must be why samtools does not want to read a bam file with suppressed fields... But then I am really confused as to how one is expected to provide a bam file to riboseqr...

[GenoMax]

Post process normal file to get what riboseqr needs?