Closed TuomasBorman closed 4 months ago
One option is to catch na.rm
in .agg_for_prevalence
function and create new parameter called drop.empty.rank
--> don't feed it to agglomreateByRank
function instead use drop.empty.rank
--> Use na.rm in rowSums in line 216
Yes the drop.empty.rank
might be more clear.
TBH I am not sure how necessary it is to allow users decide whether they want to have to have na.rm
TRUE of FALSE in the merging. But perhaps it is good to let user decide ultimately.
I would pay some attention for documentation, examples & unit tests to keep it clear for users.
Is it now good to go?
If there are NA values, they are interpreted as FALSE i.e., certain value do not exceed the detection threshold i.e. "this taxon cannot be found from this sample"