Singleton RNA on DNA trio fallback is a bit dangerous

[dnil]

Describe the bug
When uploading a WTS outlier trio file on a DNA singleton case, the sample ids appear to fall back to the singleton individual.
Trio-on-trio upload, single-on-single are ok, as is RNA singleton on DNA trio (in which case the fallback is appropriate).

See RNA case funnyhookworm, trio on trioeagerdingo and in particular RNA trio on muscle DNA singleton poeticiguana. In the latter, all variants appear to originate in the affected individual. Contrast eagerdingo where only 8 are outliers in the affected.

[dnil]

Well, after some detective work I would like to step back on the "bug" label here: the fallback is quite explicit, so formally more of a feature. But I don't think we fully discussed this situation when thinking about loading - we were a lot more focused on the singleton-RNA on trio DNA. The problem is that again that the WTS tsv files alone do not contain the subject ids or other linking information. I'd say we are doing the best we can with the current information. Any more than that would have to require some more input, like ideally that cg uses load configs where the relations between the different sample ids could be made clear. Or possibly an additional toggle for the kind of fallback, or with an explicit mapping. Let's check back with what the future plans on cg/Tomte are! See discussion here Clinical-Genomics/cg#3669