Documentation or examples for chunking/splitting and gathering bam files ?

[colindaven]

Dear all,

Nextflow has great docs and patterns for splitting and gathering fasta, fastq, csv etc.

The only thing missing is the probably central file in bioinformatics (at least for alignment and variant calling), the bam file. Has anyone found any simple pipeline or example which does this easily ? Or does everyone just chunk fastqs and feed them through the pipeline in nextflow-world ?

Thanks,
Colin

[pditommaso]

So far, there's no native support. Tho we were thinking to include Biojava to implement this via a plugin. Tagging @pabloaledo

[abhi18av]

Mm, interesting point. I did some digging into this earlier and seems like the core concern of the biojava-alignment package is to basically do the alignment using different algorithms https://github.com/biojava/biojava/tree/264ebe14ce88c28c432bc76bfa6c77696f79ba1f/biojava-alignment/src/main/java/org/biojava/nbio/alignment

Maybe in the case for a BAM, it's worth considering the htsjdk library since it has a BAM file reader https://github.com/samtools/htsjdk/blob/master/src/main/java/htsjdk/samtools/BAMFileReader.java and offers support for effective reads/writes for other file formats such as CRAM / VCF etc

Thoughts?

[pditommaso]

Abinhav, you are right. Actually, I remembered @lindenb once made an implementation for this. We could recover that code and turn it into a plugin.

[lindenb]

Abinhav, you are right. Actually, I remembered @lindenb once made an implementation for this. We could recover that code and turn it into a plugin.

at the end, I wrote a the samtools subcommand samtools samples . I now usually split the output of this command using a simple Channel.spitCsv

For more complex queries, I usually write some small htsjdk-based java codes in a process in a "Here Doc".