error running RunQuantileAlignSNF on Seurat object about liger HOT 5 CLOSED

This error message means that there are some cells for which all H factor loading values are 0. This is usually caused by the presence of some cells that have no expression for any of the selected genes. Try checking your seurat object to see if there are any cells with H factor loading values that are all 0, and if so, remove any such cells.

from liger.

Also, I don't recommend "regressing out" anything before running LIGER. One of the advantages of our method is that technical factors (mitochondrial percentage, batch, etc.) often get represented as individual factors. You can then remove them simply by not using those factors in downstream analyses. Check out the recent issue on this topic for more info.

from liger.

In your Seurat object, the H factor loadings should be available through seurat_obj[['iNMF_raw']]@cell.embeddings -- you can also see how where other components of the factorization are stored in the SeuratWrappers code.

from liger.

Thank you very much for your comments, Joshua!

Could you please instruct how to find out H factor loading values for each cells in a Seurat object? I happened to read your other post using something like:
cells.use = c(names(which(rowSums(liger_obj@H[[1]][,use_these_factors]>0)>0))

Is there a place in Seurat object that store the H factor matrix after running the RunOptimizeALS command?

Thank you!

from liger.

Thank you! Now I can find it!

from liger.