is sizeFactorNames being removed now in scran?

[LuyiTian]

sce = computeSumFactors(sce)
sce = normalize(sce)
var.fit = trendVar(sce, method="loess", use.spikes=TRUE)

you will have errors:

Error in sizeFactorNames(object) : could not find function "sizeFactorNames"

in normalize, add centre_size_factors = FALSE avoid the error but you still have the same error when running trendVar. I think it is because trendVar check the size factors by .check_centered_SF(x, assay.type=assay.type).

[LTLA]

IIRC, sizeFactorNames was added to the BioC-devel version of SingleCellExperiment. So if you only have the BioC-release version of SingleCellExperiment installed, it probably won't work.

[LuyiTian]

Ok, think you. I checked scran and scater but havnt think about SCE. I installed the new one because the current release has a bug in MDS that it gives the same result even when you use different expression slot assay.

BTW. My benchmark dataset shows your scran is the best in normalizations, better than these methods that have complicated model.

[LTLA]

Thanks for letting me know, that's reassuring.

[dhuvik]

Hi Aaron, I seem to be having the same problem but on the current BioConductor (3.6) It wont update to 3.7 using biocLite() for some reason. However, when trying to use the calcAverage function I get an error saying could not find function sizeFactorNames. https://support.bioconductor.org/p/109555/ and https://stackoverflow.com/questions/50613564/could-not-find-function-sizefactornames-bioconductor-r3-5 for your reference. I have tried reverting back to the non-devel versions of the packages as well with no success

[LTLA]

Well, fix your update problem first; read https://bioconductor.org/install/#troubleshoot-bioconductor-packages.

[dhuvik]

Thank you! Removing all packages installed through BioConductor and Reinstalling BioConductor seemed to fix everything.