Please Explain "Fold Enrichment" In Useq'S "Intersectregions" Tool

Question

[Atom Smasher]

Hello,

I was trying to use USEQ's "IntersectRegions" tool to find genomic overlaps. However, I don't think I have understood the concept of "Fold Enrichment" very well. What exactly is "fold enrichment" ? I am aware that it's some kind of "signal over noise" but I'll appreciate it if I can get an explanation of it in simple language.

Also, to calculate fold enrichment, USEQ requires me to use a parameter -r .

They describe it here as follows:

-r Make random regions matched to the second regions file(s) and intersect with the first.

I did not quite understand what "make random regions matched to the second regions file" means.

It'll be great if I could understand this.

Answer 1

[Lars Juhl Jensen]

Fold enrichment is a general statistical term not in any way specific to USEQ. It simply means how many fold more did something happen than you would expect by random chance, that is a fold enrichment of 2 means it happened twice as much as your random expectation.

I do not know USEQ, but I assume that the -r option produces a large number of random regions that have statistical properties similar to the regions you provide in the second file. It then does the analysis using both the real regions from the second file and the random regions. Finally it compares the overlaps found for the real regions to those found for the random regions and based on that calculates, if what you saw for the real regions is more than what you get for random regions. This result is probably summarized as a p-value and a fold enrichment.

What I outline above is the general way for calculating p-values and enrichments by stochastic simulation. It is the most useful thing I ever learned in a statistics course!

[Atom Smasher]

Thank you very much. That was very nicely explained :)