feature request: Make leafcutter agnostic of chromosome identifiers about leafcutter HOT 6 CLOSED

On Fri, Oct 27, 2017 at 10:14 AM, shoersch ***@***.***> wrote: Hi, It would be fantastic - essential for me ;) - if leafcutter could process non-human or non-standard chromosome identifiers. Currently, it cannot, as per this line in leafcutter_cluster.py: chromLst = ["chr%d"%x for x in range(1,23)]+['chrX','chrY']+["%d"%x for x in range(1,23)]+['X','Y'] and skipping everything else entirely ( if chrom not in chromLst: continue ). Use cases for this request abound and include: - organisms with chromosome counts exceeding the number 23 (zebrafish, dog, cow...) - organisms with different chr. naming schemes (cat, fruit fly, nematode) - unplaced scaffolds, alternate versions (haplotypes) - alignments against genome assemblies using different identifiers (e.g. NC_000075.6 via NCBI) Arguably, leafcutter should make no assumptions whatsoever as to how chromosomes are identified. Chromosome IDs could be provided by the user in a file, or (much preferred!) be generated on the fly based on a digest of the .junc files generated before, as per a preprocessing step based on something like: cut -f1 *.junc | sort | uniq Thanks much, Sebastian PS: Related to issue #25 <#25> — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#42>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AE1piXsMd4Ggl4ZFO4Ix9ayde8OCKxy3ks5swfNwgaJpZM4QJNSJ> .

On Fri, Oct 27, 2017 at 11:28 AM, shoersch ***@***.***> wrote: Thanks much - so, commenting out instances of "chromLst = " and corresponding "if chrom not in chromLst:" statements (2 instances, I think - in "pool_junc_reads" and in "sort_junctions" ), allows the script to run, apparently ok (although I haven't evaluated thoroughly!), now producing clusters with the non-standard chr IDs etc. Could this, then, be made a user parameter? For now, I can live with this hack, but I would be very concerned, if I had to re-hack this script every time a new version of leafcutter comes around... Sebastian — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#42 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AE1piXrT0i7kKPYXVWSHLFn_NEfOgP6lks5swgSzgaJpZM4QJNSJ> .