Closed halterc closed 11 months ago
Good suggestion, this has now been implemented in the latest version of ProjecTILs.
You should now be able to run ProjecTILs.classifier()
either on a single object (optionally splitting by a metadata column) or on a list of objects. In the latter case, parallel processing is also available (ncores
parameter) for the analysis of multiple samples in parallel.
-m
Hi, in contrast to Run.ProjecTILs, which runs just fine taking as input a list of seurat objects, ProjecTILs.classifier returns an error: Error in ProjecTILs.classifier(test, map, ncores = 3) : Query must be a single Seurat object
EDIT: Instead of supplying a list of seurat objects directly, supplying a single seurat object to ProjecTILs.classifier and supplying the "split.by" parameter (e.g. split object by sample into list of objects) provides a workaround.