Conversation
|
Hi, Not member of dev team but just a thought here. A check should also probably be added to prevent integrated assay from being selected here because as you mentioned it doesn't have counts and shouldn't be used, would just provide more informative error message to end user. Best, |
|
Hi Sam. It is a good point. As of now it runs without errors, but just returns I wouldn't be confident throwing errors from functions that currently return. Might mess up something elsewhere. |
This is meant to help diagnose issues when assay doesn't have desired layer. For example see satijalab/seurat#8658 and discussion in satijalab/seurat#8659. The warning message might not be the prettiest, but I didn't want to go through `Layer.Seurat` method to get the assay name.
|
I'll leave up to dev team to decide but personally I think abort in percentagefeatureset makes more sense to me. End users may still end up confused (for those that don't understand integration doesn't have corrected counts) as to why function fails. And a warning about integrated assay not being appropriate for this particular function makes sense. But either way works. Best, |
This is meant to help diagnose issues when assay doesn't have desired layer. For example see satijalab/seurat#8658 and discussion in satijalab/seurat#8659. The warning message might not be the prettiest, but I didn't want to go through `Layer.Seurat` method to get the assay name.
Fixes #8658