Simplifying WDL workflow submission

[tefirman]

Description

• In an effort to make our WILDS WDL's more accessible to HPC-curious researchers, we are making the input json as minimal as possible:
  • All inputs other than fastq locations are now optional.
  • If a reference fasta & gtf are provided, the script will use those, but if not, it will download GRCh38 from NCBI.
  • To avoid confusion, the script will automatically build a STAR reference directory from the reference fasta/gtf.
  • To avoid data duplication, it is highly recommended this workflow be run with call caching enabled.

Related Issue

• Fixes Unclear where to find input reference files #5

Testing

• Submitted minimal input json via PROOF, works as expected.
• Meeting with user on 12/5 to see what they think.

[tefirman]

Short-staffed at the moment and need this WDL for a user study. Merging without additional review.