Closed jolespin closed 6 years ago
Hi - re-running your code it highlights that there are indeed 32 samples without any positive OTUs.
cumNormStat
and cumNormStatFast
will not work if there are no positive OTUs.
sum(colSums(obj_f)==0) [1] 32
Shouldn't idx_samples_f = which(colSums(obj) > 0)
make sure that only samples with a library size greater than 0 ?
Like on page 10:
It's your ordering. You're calculating the column sums on pre-filtered data.
I'm trying to follow a slight variant of the tutorial to get started from ~ page 9: https://www.bioconductor.org/packages/devel/bioc/vignettes/metagenomeSeq/inst/doc/metagenomeSeq.pdf
Essentially, I need to figure out how to go from: (A) my raw Otu counts; and (B) my "class" assignment (e.g. diseased, healthy) to get (C) differentially abundant Otus.
I'm starting with the built in dataset lungData but I'm getting an error when I'm calculating the normalization factors.
I've made sure there are no empty features (Otus) or empty samples. I've documented my code heavily to serve as a tutorial in case anyone needs to learn how to use the pipeline.
I'm looking at the source code and this is the condition my data is breaking:
Here is my code: