Filtering values for two group separately

Question

[Expert]

There are 6 controls and 6 treatment variables in a dataset. I used these command first:

mydata <- read.delim(".csv", header = TRUE, sep=",", stringsAsFactors = FALSE, row.names = 1)   
targets <- data.frame(fileName=colnames(mydata)[1:12], Group=rep(c("Control" , "Treatment"), each= 6))

I want to filter for remove low counts, But in the same table this filtering will be:

control of %20 <1 , treatment of %20 <1 separately.

I could not filter in the same table.

Can you help me? Thank you.

[Asaf]

A. You won't get a reply when people don't understand what you're asking B. Don't use FPKM

[WouterDeCoster]

How is at least twenty percent of each group is filtered to be greater than zero? That is 20% of the control group and 20% of the treatment group separately.

That sentence as it is written doesn't make sense at all. Please elaborate, and it may be helpful to include examples.

Answer 1

[Kristoffer Vitting-Seerup]

You should not filter individually in each group! Filtering in individual groups will inflate the false positive rate on any statistics you do downstream of the filtering.

[Asaf]

You should probably not filter at all