I have a problem in which I am trying to use the zinbwave R package for differential expression analysis from snRNAseq data, but I cannot get the necessary observational weights.
I am using: R version 3.6.1 zinbwave_1.8.0
I created a SummarizedExperiment object by using data and metadata from a Seurat object:
se <- SummarizedExperiment(seurat@assays$RNA@counts, colData = email@example.com)
After filtering the se_object by gene counts, I use the 'zinbwave' command to generate a SingleCellObject
singlecell <- zinbwave(se, K=2, X="~genotype", epsilon=1000)
The SingleCellExperiment object ('singlecell') is generated, but after creating the object there is no "weights" data in the object.
Trying to access weights data:
weights <- assay(singlecell, "weights")
results in this error message:
Error in assay(singlecell, "weights") : 'assay(<SingleCellExperiment>, i="character", ...)' invalid subscript 'i' 'weights' not in names(assays(<SingleCellExperiment>))
In the documentation, it says that "Since version 1.1.5, zinbwave computes the observational weights by default." However, I also tried specifying "observationalWeights = TRUE" and this did not appear to lead to any weights being generated, either.
Why are the weights not generated by the zinbwave function? Could it be because of a problem with how I set up the SummarizedExperiment? Or is it more likely related to the way I have used the zinbwave command?