peak to gene

[rojinsafavi]

I have a question regarding the peak to gene correlation analysis.
Based on my understanding it is possible to find a peak that is correlated a gene, but the TFBS in that peak has no fragment aligned to, basically, it is not accessible, similar to the following case:

Doesn't that make the peak to gene correlation a bit misleading? What if I want to know where in the peak(500bp) there is accessibility exactly?

[rcorces]

This seems like a pretty unlikely situation given that most peaks would require many fragments in order to be called a peak. That being said, a few comments:

1. a peak to gene link is really just a correlation and doesnt imply that any specific TF is actually causing a relationship
2. because of how insertions happen, the two insertions of a fragment dont necessarily need to span a TF binding site

What if I want to know where in the peak(500bp) there is accessibility exactly?

This would be what single-site footprinting would be good for but I dont personally think scATAC-seq is a great way to get at this question.