enhancer width

[ferenckata]

Hei,

I was using CAGEfightR enhancer calling (quickEnhancers) through CAGEr and for curiosity plotted the width of the enhancers. I can see that in all of my samples the enhancer widths show a bimodal distribution (see below) and I wonder why. Do you have an explanation maybe?

[MalteThodberg]

Hello @ferenckata ,

CAGEfightR will sometimes merge nearby bidirectional clusters if they are very close: I suspect that is what you are seeing here.

To investigate what is going on, you could use trackBalance to plot the balance score for one of the wide enhancers.

[ferenckata]

Thanks, I will look into that!

Is there any way I can disable this merging functionality?

[MalteThodberg]

(late answer, but in case you are still interested).

At the moment the merging of bidirectional peaks is hard coded. This is to prevent having overlapping clusters, which messes with quantification and differential expression later.

You could directly modify the code for bidirectional clustering without too much trouble though by stopping before the reduce step: https://github.com/MalteThodberg/CAGEfightR/blob/master/R/enhancers.R