Error in tcrossprod(model_pars_final[genes_bin, -1, drop = FALSE], regressor_data_final) : "tcrossprod" is not a BUILTIN function

[Counts-Xin]

cso<-SCTransform(cso)
Calculating cell attributes from input UMI matrix: log_umi
Variance stabilizing transformation of count matrix of size 23435 by 65231
Model formula is y ~ log_umi
Get Negative Binomial regression parameters per gene
Using 2000 genes, 65231 cells
|======================================================================| 100%
Found 88 outliers - those will be ignored in fitting/regularization step

Second step: Get residuals using fitted parameters for 23435 genes
| | 0%Error in tcrossprod(model_pars_final[genes_bin, -1, drop = FALSE], regressor_data_final) :
"tcrossprod" is not a BUILTIN function

[saketkc]

Can you paste your sessionInfo()?

Alternatively, can you reinstall Matrix and sctransform and try again?

remove.packages(c("Matrix", "sctransform"))
install.packages("sctransform")

[Counts-Xin]

Since I was more familiar with Seurat v3, I installed Seurat v3 on R 4.x to maintain compatibility, because DoubletFinder requires R 4.x. I have tried multiple versions of sctransform (including 0.3.2, 0.3.5, 0.4.1, and 0.4.2) as well as Matrix (versions 1.6-4 and 1.3-2), but the issue persists.