Error in creating grouped matrices

[ccruizm]

Good day!

I would like to implement this nice package on our cancer data. I have tried but ran into and error in the first step:

creating group matrices

Warning message:
"The `slot` argument of `GetAssayData()` is deprecated as of SeuratObject 5.0.0.
ℹ Please use the `layer` argument instead."
Warning message in asMethod(object):
"sparse->dense coercion: allocating vector of size 237.8 GiB"
computing epiCHAOS scores

calculating pairwise distances for group-MP 2 chr1:

calculating pairwise distances for group-MP 13 chr1:

calculating pairwise distances for group-MP 7 chr1:

calculating pairwise distances for group-MP 9 chr1:

calculating pairwise distances for group-MP 3 chr1:

calculating pairwise distances for group-MP 12 chr1:

calculating pairwise distances for group-MP 5 chr1:

calculating pairwise distances for group-MP 8 chr1:

calculating pairwise distances for group-MP 11 chr1:

calculating pairwise distances for group-MP 4 chr1:

Error in lm.fit(x, y, offset = offset, singular.ok = singular.ok, ...): 0 (non-NA) cases
Traceback:

1. epiCHAOS(counts = GetAssayData(malignant_atac, slot = "counts"), 
 .     meta = malignant_atac@meta.data, colname = "MP", n = 500, 
 .     index = NULL, plot = TRUE, cancer = TRUE, subsample = 1)
2. compute_eITH.cancer(x = matrices)
3. lm(unlist(dists) ~ unlist(coverage))
4. lm.fit(x, y, offset = offset, singular.ok = singular.ok, ...)
5. stop("0 (non-NA) cases")

What do you think could be the cause?

[quail768]

Hey @ccruizm
Can you check if your row names in the count matrix are chr1:181291-181562 and not chr1-181291-181562 . This was causing the same error for me and the program works without issues after fixing.

[debbie28]

Hi! Not from the dev team, but based on your description it seems you might have put cancer=T. I got the same error, which was fixed when I put cancer=F.

[quail768]

Hi! Not from the dev team, but based on your description it seems you might have put cancer=T. I got the same error, which was fixed when I put cancer=F.

Yes @debbie28 it looks like the functions for cancer=T and cancer=F are different because of the per-chromosome count correction that is applied for cancer datasets. If cancer=F then it looks like count correction is not applied.