Lightning

Take code in https://github.com/curoverse/l7g/tree/master/tools/tagset and convert it to a CWL pipeline for automatic tagset creation.

This includes:

• Create a FASTA file that has the tagset (what the referenced code does now)
• Compress and index (both for samtools faidx access and bgzip access) the created FASTA tagset
• Create a 2bit tagset

The original code to create the tagset was created using Go(lang) and has since been ported (in the above code referenced) to a combination of Python and C++.

Assuming all dependencies have been met, simply running createTagsetFa.sh should work to output the FASTA tagset.

Dependencies include:

• bigWigToBedGraph (available at http://hgdownload.soe.ucsc.edu/admin/exe/linux.x86_64/)
• hg19 FASTA file (available at http://hgdownload.cse.ucsc.edu/goldenPath/hg19/bigZips/ ?)
• cytoBand.txt (available at http://hgdownload.cse.ucsc.edu/goldenPath/hg19/database/)
• wgEncodeCrgMapabilityAlign24mer.bigWig (available at http://hgdownload.cse.ucsc.edu/goldenPath/hg19/encodeDCC/wgEncodeMapability/)

All these files are available in Keep via the Project reference.

The original tagset creation had some issues including not always falling on a unique 24mer and some inconsistencies in choosing an end tile length and tag. The quirks of the original tagset creation have been encoded in the choose_tagset_startpos0_vestigial.py script. In the future something like choose_tagset_startpos0.py or another custom script could be used to choose a tagset.

Though outside of the scope of this work, it would be nice to come up with a CWL pipeline that somehow made custom tagsets easier to experiment with. Converting the current tagset creation into a CWL will hopefully be a good start.