CreateSeuratObject taking very long

Question

[eae6d2e7]

I have my data with 33694 obs of 63690 variables, and it has been 2 hours since I ran the below command and it still isn't complete.

seu_obj<-CreateSeuratObject(count=raw_data)

Is there any way to speed this up?

Also I am getting this warning message "Warning: Data is of class data.frame. Coercing to dgCMatrix." Is this the reason for long time?

Posted here as well https://www.reddit.com/r/bioinformatics/comments/1cwfx59/createseuratobject_taking_very_long/

[Nitin Narwade]

I am not sure whether it will speedup the conversion but you can give it a try. convert your dataframe into a sparse matrix before creating a seurat object.

raw_data <- Matrix::sparseMatrix(raw_data)
seu_obj<-CreateSeuratObject(count=raw_data)

Side not, check for duplicate entries. I am not expecting this will speed up the conversion or anything but it is just a good practice to look at it in advance.

All the best.

Regards, Nitin N.

[eae6d2e7]

Thank you for your reply!

I followed your advice to convert to sparse matrix first but I got this error Error in Matrix::sparseMatrix(raw_data) : exactly one of 'i', 'j', and 'p' must be missing from call

What does it mean?