Might want to attach WDL running to the SnakeMake to fill in the calling results automatically.
Statistics:
- Mapping speed
- Softclipped and unmapped bases
- Simulated mapping accuracy
- Variant calling accuracy
Conditions:
- Mapping to HPRC M/C graph with Giraffe (minimum condition for all techs to say "it works" and fix a version)
- Might need to add next pangenome release in before paper comes out
- Mapping to linear stick graph with Giraffe
- Mapping to 64-haplotype downsampled 1KG graph from original Giraffe paper
- Mapping to linear reference with Minimap2 (but also BWA-MEM for Illumina)
- Maybe mapping to a more collapsed graph as an internal metric
Read Sets:
- HiFi
- R10
- Illumina (paired end) (non-chaining codepath) (haplotype sampling mode only)
- Lets us say we have one mapper for long and short reads even though it's a different codepath, "universal mapper"
Could be a big Google Sheets.