Cholmod error 'problem too large'

[sciotlos]

Hi all,

I'm getting the following error when trying to create the Tempora object with ImportSeuratObject:

Error in asMethod(object) : 
  Cholmod error 'problem too large' at file ../Core/cholmod_dense.c, line 105

The dimensions of the Seurat object are as follows:

> subset_merge_allCells
An object of class Seurat 
60218 features across 102868 samples within 2 assays 
Active assay: SCT (26680 features, 3000 variable features)
 1 other assay present: RNA
 2 dimensional reductions calculated: pca, umap

I know this is a large object, I've actually cut it down in half (I have a little over 200k cells sequenced). And so I wonder if anyone has some suggestions for:

1. stripping down the data further - could I get rid of all genes except the 3000 most variable ones?
2. remove more cells during the prelim QC of cells? I've already halved the dataset, so I'd rather not do this one..
3. is there a general dimension/size limit for creating the tempora object?

Thank you!

[sciotlos]

quick update - I was able to get around this issue by removing genes expressed in less than a certain number of cells. The new object is 12355 features x 102868 cells, and ImportSeuratObject does not complain.

I'll leave this issue open for now, in case someone can reply with more specific limitations for the importseuratobject function.

[zh-zhang1984]

I also encounter the same problem with the "RNA" slot in seurat object; however, when I use the "integreated" assay, which contained only 2000 variable RNAs, the problem solved.