Feature request: Option for unsorted BAM output by dorado aligner

[cihanerkut]

Dorado aligner output is sorted by default (#642) but some tools may require alignments to be at random order, i.e. unsorted. As far as I could see, this is the case when the output is directed to stdout rather than an output folder:

dorado/dorado/cli/aligner.cpp

Lines 276 to 277 in 9dc15a8

	const bool sort_bam = (file_info.output_mode == utils::HtsFile::OutputMode::BAM &&
	file_info.output != "-");

However, the alignment_summary.txt file also has valuable information but it won't be created unless the output is directed to a folder. To break this mutual exclusivity, one can perhaps add a command line argument to disable sorting alignments, basically overriding the code above.

My current workaround is using samtools collate to "shuffle" the alignments but it means unnecessary compute.

[susie-ont]

Hi @cihanerkut, thanks for your post. To help us better understand your request, could you please specify which tools you’re using that require shuffled alignments?

[cihanerkut]

Hi @susie-ont,

Thanks for your quick response. I would like to use Salmon, for example. Please check the note "Read / alignment order" in the manual:

https://salmon.readthedocs.io/en/latest/salmon.html#using-salmon

An alternative solution could be to allow dorado aligner to write alignment_summary.txt in an arbitrary path provided as a command line argument and BAM output sent to stdout, unsorted.