Changes in vignette on "Perform an integrated analysis" #4017
Replies: 3 comments 4 replies
-
Hi Peter, Can you elaborate a bit on the differences are you seeing in terms of performance? From what I understand you are asking about the difference between: features <- SelectIntegrationFeatures(object.list = object.list)
anchors <- FindIntegrationAnchors(object.list = object.list, anchor.features = features) and anchors <- FindIntegrationAnchors(object.list = object.list) In this situation, there shouldn't be any difference in terms of performance. The first chunk is just slightly more explicit about which features are being used. In |
Beta Was this translation helpful? Give feedback.
-
Hi Andrew |
Beta Was this translation helpful? Give feedback.
-
hi,
thank you for your help! |
Beta Was this translation helpful? Give feedback.
-
Hi SatijaLab
After the recent update of your website, I have noticed that you changed the pibeline for "Perform an integrated analysis" using the NormalizeData + FindVariableFeatures preprocessing. In the old pibeline you did not use the "SelectIntegrationFeatures" before finding anchors, but instead created a refrence.list containing the datasets, which you used as input for FindAnchors + an arguments for dimensions.
Could you maybe explain the difference between these two approaches in the detection of anchors (Using SelectIntegrationFeatures vs. reference.list)?
I would really appreciate this as it seems to make a significant difference in the performance
Best regards,
Peter
University of Southern Denmark
Beta Was this translation helpful? Give feedback.
All reactions