Closed LeeTL1220 closed 2 years ago
(With the latest datasources, esp. gnomAD 3.x)
Given the large volume of data needed to run Nirvana (about 20GB), creating a performant docker image is not feasible. Therefore, we have chosen not to publish docker images for Nirvana.
Huh? The docker image for NIRVANA should not include datasources -- those get mounted externally. The image for 3.14 didn't include the datasources and it was quite performant. Why would 3.18 be any different?
I just meant that the Downloader in the docker image should download the latest version of the datasources.
@rajatshuvro
(Not to beat a dead horse but...)
To give you an idea of scale, we used NIRVANA 3.14 in a docker image (that you provided) to annotate >700M variants for our 98k sample 30x WGS joint callset. NIRVANA was performant.
@RoriCremer FYI
Can you create a docker image and tag (in dockerhub) for the latest version of NIRVANA?