footprint bw track

[YichaoOU]

Hello,

There are 3 bw files from pred_bw.

bias.bw
chrombpnet.bw
chrombpnet_nobias.bw

I suppose chrombpnet.bw is for footprint track?

Thanks, Yichao

[panushri25]

If you want a track to compare with the observed tracks directly you will use - chrombpnet.bw (which is not corrected for bias)

But if you want something which is corrected for bias and reflects the footprints of TFs - please use chrombpnet_nobias.bw

Hope this answers your question, otherwise please post back.

I will improve the documentation to clarify this, thank you!

[YichaoOU]

Thanks! Below is a screenshot for 3 bw tracks for one sample (same sample).

chrombpnet_nobias.bw

chrombpnet.bw

bias.bw

It seems the chrombpnet.bw is in high resolution, which looks more like a footprint track? I want to find some differential footprints comparing different conditions.

[panushri25]

What do you mean by high resolution?

[panushri25]

Can you lay it over with the observed bw

[YichaoOU]

High resolution. For example, the last track is a footprint track using HINT-ATAC (different cell, as compared with the above chrombp tracks). You can see for HINT-ATAC footprint track is more similar (in terms of the style) to chrombpnet.bw than chrombpnet_nobias.bw.

I will add observed bw later. Thanks for the quick reply!

[panushri25]

If you add the observed track you will be able to tell the difference yourself i think. I suggest getting contribution score bigwigs for some of these regions to get some sense, you will see that chrombpnet_nobias.bw will directly overlay with motifs and show footprints.

Are you referring to the spikiness as high resolution? I am a little confused can you elaborate?