Hi! 

I am happy with the results given by filtering with prior information , but it will be actually nice to filter a bit by abundance those regions. I read in the documentation that it is possible , but I am struggling to make it work. Should I implement the aveLogCPM into overlapsAny function? I am bit lost, sorry.

Thanks in advance

Sergio 

Let's say you've got a logical vector after filtering your windows on abundance. For example, if you have a SummarizedExperiment object named data:

ab <- aveLogCPM(asDGEList(data))
keep.ab <- ab > 0

Of course, you could use any of the other filtering strategies that are listed in chapter 3 of the user's guide, so long as you get a logical vector with length equal to the number of windows in data.

Now, let's say you want to filter to keep windows in a bunch of regions as well.

keep.reg <- overlapsAny(rowRanges(data), regions)

You can then combine the two filter vectors with a simple AND operation (see Section 3.7 of the guide). This will retain only those windows that are high-abundance and lie within one of the specified regions:

keep <- keep.ab & keep.reg
filtered.data <- data[keep,]

Of course, this is more stringent than filtering on abundance alone. Make sure that you have enough windows for stable calculation of downstream statistics (at least several thousand, usually).

Ok, perfect!!! Thanks I a lot , I will try! 

Good, it is exactly what I needed. Thanks a lot. 

One more question Aaron. When I predefined the regions , for example promoters, is there a way to avoid the counting per window and just count per predefined region? Maybe it is actually what I am doing, sorry if I misunderstood it. 

thanks a lot! 

Sergio