EDIT by RamRS:

Cross-posted on bioinfo SE: https://bioinformatics.stackexchange.com/questions/21455/try-to-understand-better-how-to-set-out-an-effective-sample-size-within-the-meta

Dear all,

I am trying to understand better what number to be used and how to put it in the function MRcoefs within the Metagenomeseq package in R for calculating differentially expressed OTUs between two condition:

My code looks like:

MRcoefs(fit, coef = 2, group = 3, number = 143, eff = 5.5) 

Here I am setting the eff argument to 5.5 because I have calculated the effectiv_size using this code:

Effective_size(fit) 

which result in table of effective size for each feature (OTU), and then I got the mean of these

mean(Effective_size(fit))

which was 5.5

When I run this, it produced an error: prop outside [0,1], I understand it should be between 0-1, but why ? is not that number in eff is the mean that I calculated which bring 5.5 ? Also when I tried to run it with .5 as the document suggested as below :

MRcoefs(fit, coef = 2, group = 3, number = 143, eff = .5) 

it run nicely, no error , but it produced the same OTUs as when I did not include eff, I was expecting that some OTUs will be removed as they were below this threshold

Could you please help me?