1
$\begingroup$

I need to perform differential expression analysis using the scDD package from R, but I am not able to since I miss the normcounts assay in my SCE object (of course in the example they show, the assay is already there).

As I found out in the preprocessing steps, one can use this function to generate the missing assay, except that it does not work:

# this is the SCE object
> cTSubSet
class: SingleCellExperiment 
dim: 19767 149 
metadata(2): Samples Samples
assays(2): counts logcounts
rownames(19767): Xkr4 Gm19938 ... CAAA01147332.1 AC149090.1
rowData names(4): ID Symbol Type chrLoc
colnames(149): AAAGGATGTTCACGAT-1 AACCATGTCTTTCGAT-1 ...
  TGTGGCGTCTCTTAAC-1 TTTGGTTAGACGCTCC-1
colData names(17): Sample Barcode ... slm_1 slm_1.5
reducedDimNames(0):
mainExpName: NULL
altExpNames(0):

# setting parameters as suggested in the vignette linked above
> paR <- list(alpha=0.01, mu0=0, s0=0.01, a0=0.01, b0=0.01)
# setting zero.thresh to zero since I already removed unwanted genes
> cTSubSet <- preprocess(cTSubSet, zero.thresh=0, scran_norm=TRUE)

# the not so useful error:
Performing scran Normalization
Error in t.default(Data) : argument is not a matrix

I was also looking for other possible ways of obtaining such normcounts assay and I found this description:

normcounts: Normalized values on the same scale as the original counts. For example, counts divided by cell-specific size factors that are centred at unity.

Still, I can't figure out what this may mean. Is there a way of obtaining such normcounts?

Improve this question
$\endgroup$
2
  • $\begingroup$ You might as well try monocle3::normalized_count. It takes SingleCellExperiment object. $\endgroup$
    – littleworth
    Commented Jun 17, 2021 at 10:31
  • $\begingroup$ yep that is a possibility. I actually found another one using edgeR, i.e. y <- DGEList(counts=(counts(cTSubSet)), group=cTSubSet$condition) y <- calcNormFactors(y) however, I am still working on this $\endgroup$
    – gabt
    Commented Jun 18, 2021 at 12:01

1 Answer 1

Reset to default
2
$\begingroup$

I do not know this package but it seems to call scran so why bothering and not call scran directly, see for an extended discussion + code:

https://bioconductor.org/books/release/OSCA/normalization.html#normalization-by-deconvolution

Improve this answer
$\endgroup$
1
  • $\begingroup$ yes, I am familiar with the OSCA pipeline. At this point I am just wondering if the values that are returned by computeSumFactors() can be considered as normcounts. $\endgroup$
    – gabt
    Commented Jun 21, 2021 at 7:24

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.