BAM vs fastq input

[David Minkley]

Hello,

I was wondering what sort of actions abyss-pe performs on BAM files when used as inputs, as opposed to fastq.  Until this point, I've been using exclusively fastq data (since I've already converted it to fastq for use with other programs).  Does abyss use the BAM files directly, or does it first preprocess them (and convert them to fastq)?  The reason I ask is that in allpaths-lg there is a time-consuming pre-processing step if you use BAM data which isn't present for fastq.

Thanks!

- David

[Tony Raymond]

Hi David,

Any time ABySS reads a BAM file, it creates a 'samtools view -h' process that will pipe the converted SAM output into whatever application is reading the file. In this way, reading BAM files is not a separate step, abyss will not use any additional disk space, and it shouldn't take any longer than reading a fastq file.

This is also how ABySS reads compressed files.

Cheers,
Tony

[David Minkley]

I finished the assembly overnight, using bam input files.  The resulting file is a *-unitigs.fa file, rather than the *-contigs.fa file I'd expect (the BAM file contains paired data).  How do I force ABySS to treat the data within the BAM file as paired-end?

My abyss run command was along the lines of:

abyss-pe np=43 k=81 n=10 name=abyss_genome_k81_n10 lib='lib01 lib02'\
 lib01='lib01.bam'\
 lib02='lib02.bam'\

-David

[Tony Raymond]

Hi David,

I don't see any problem with your command line. Was there an error in the logs? If you use the v=-v parameter of abyss-pe, abyss will print some additional debugging information that might be helpful.

If the problem isn't obvious from the logs, would you send me the stdout and stderr log files from abyss-pe?

Thanks,

Tony