By using JASPAR API and PWMScan which tool is more accuracy to predict transcription factor binding sites

Question

[michaelchen33]

As title.

I am using JASPAR data: MA0058.2(transcription factor: MAX) to scan whole genome(hg19), I found out that JASPAR has already did the job and put the binding site data in the bed_files/ (http://jaspar.binf.ku.dk/html/DOWNLOAD/), there are about 24564 binding sites that it found by using MA0058.2 to scan hg19.

When I use PWMScan (http://ccg.vital-it.ch/pwmscan/) to scan the same data (MA0058.2) in hg19, I got about 131785 binding sites that it found.

I am curious why there is such big range in them?

Which tool is more accuracy or any suggestion tools that I should try?

Answer 1

[asntech]

Hi,

The bed files in JASPAR are the sites, which are used to construct the position frequency matrix (PFM), not the genome-wide TF binding sites. The 2018 release of JASPAR now provides predicted genome-wide TF binding sites (TFBSs) for several species as genome tracks and also bed files. More details are available here http://jaspar.genereg.net/genome-tracks/

P.S. Although, it is a late answer, but I'm posting this if someone else have similar questions.