Yeah it would be awesome to make it a meta pipelin...
# benchmarkingn
Nicolas Vannieuwkerke
11/21/2023, 1:30 PMYeah it would be awesome to make it a meta pipeline in the future (as this should be able to run on almost all outputs of the other pipelines)
r
Rike Hanssen
11/21/2023, 1:37 PMthis sounds like the world domination pipeline, sarek has been aspiring to be 😄
🤣 2
🤌🏻 1
n
Nicolas Vannieuwkerke
11/21/2023, 1:38 PMwho knows 😉
Nicolas Vannieuwkerke
11/21/2023, 1:38 PMSarek gobbling it up lol
r
Rike Hanssen
11/21/2023, 1:39 PMI am just wondering whether this is better/worse than having several pipelines at least per domain to some extend. I guess e.g. rnaseq might need very different tools and experimental setups for benchmarking
n
Nicolas Vannieuwkerke
11/21/2023, 1:40 PMKinda true, but Sarek and raredisease for example would use mainly the same tools (and a lot of non nf-core pipelines probably too)
r
Rike Hanssen
11/21/2023, 1:41 PMyep true, thats why I was thinking domain benchmarking pipelines
j
Jose Espinosa-Carrasco
11/21/2023, 1:44 PMOur idea was to try to make something general in some dynamic manner but this probably will be difficult to achieve
n
Nicolas Vannieuwkerke
11/21/2023, 1:44 PMI agree but I'm not sure benchmarking would really be that big of a pipeline (since it's mostly one or two tools per data type)
Nicolas Vannieuwkerke
11/21/2023, 1:44 PMBut maybe we should investigate a bit more
Nicolas Vannieuwkerke
11/21/2023, 1:46 PMI was mainly thinking of creating several subwfs for all different types. Creating a samplesheet with the possibility of adding an id, vcf, tbi, and data type per sample and channel the right samples to the right subwfs (but probably easier said than done)
🙌🏻 1