search

EESI / themetagenomics

Other
23 stars 4 forks source link

Unsure why the fxn_table returns non-whole numbers #17

Open gcuster1991 opened 3 years ago

gcuster1991 commented 3 years ago

Hi,

I'm curious why the output of the t4f() function produces non-whole numbers when the input is whole numbers and no normalization is applied. Maybe i'm not understanding what the fxn_table is reporting, but I assumed it was counts of each KEGG pathway (k0001 ...) found in each sample.

taxonomy table

taxa

OTU table

seqtab.nochim

tmp <- tempdir() download_ref(tmp,reference='silva_ko',overwrite=FALSE)

FUNCTIONS <- t4f(otu_table = seqtab.nochim,rows_are_taxa=FALSE,tax_table = taxa, reference_path=tmp,type='uproc',short=TRUE, cn_normalize=FALSE,sample_normalize=FALSE,drop=TRUE) #might need unnormalized for deseq.

produces non-whole numbers.

FUNCTIONS$fxn_table

gcuster1991 commented 3 years ago

As an aside, I'm hoping to use the output of this functional assignment and import it to DESeq2. DESeq2 requires whole numbers for its negative binomial model. If another route is better, I'm happy to try that.