search

satijalab / seurat

R toolkit for single cell genomics
http://www.satijalab.org/seurat
Other
2.27k stars 910 forks source link

k.filter settings on FindIntegrationAnchors #2090

Closed DRLaks closed 4 years ago

DRLaks commented 5 years ago

Hello, What is the optimal way to set k.filter for FindIntegrationAnchors if there are less than 200 cells? Should this parameter be set as high as possible or at a fraction of total cells? The results seem sensitive to this parameter so I would like to know what is the best way to set k.filter. All the best, Dan

satijalab commented 4 years ago

Apologies for the slow reply. To be fully transparent, the Seurat integration methods (and indeed, any graph-based methods including Louvain clustering), begin to struggle with far less than 200 cells. Its difficult to give advice here, but I suppose I would personally turn off the anchor flitering (set k.filter = NA)