Confirming how to identify putative TF binding sites

[Brendan Rooney]

Hello, 

Can someone please confirm for me that I am using the tool correctly so I know my results mean what I think they do?

I have 41 differentially expressed genes from RNAseq. I want to see if there are MYC binding sites in the promoter regions of those genes. 

I am using the web interface for JASPAR. 

I added MYC to my cart, went to the cart and put the FASTA sequence for the promoter of one of my genes of interest into the scan tool and hit scan. 

Am I correct in interpreting that the output sequences are putative MYC binding sites for in the promoter of my gene?

Thanks!

[Anthony Mathelier]

Hi Brendan,

Yes, you are correct.

I would suggest that you use our set of predefined TFBSs provided as genome tracks (see https://jaspar.genereg.net/genome-tracks/).

While these predictions can be useful, they are likely to contain a high level of false positives. I would strongly recommend that you have a look at our UniBind database (https://unibind.uio.no/), which stores high-quality TF-DNA interactions with both experimental and computation evidence. We conveniently provide a TFBS extraction tool at https://unibind.uio.no/TFBS_extraction/ to make it easy to intersect your regions of interest with our sets of TFBSs. Feel free to ask if you have any questions.

Best

[Gilles Marodon]

Unibind is an amazing tool. Thank you so much for putting all this efforts in having public data reusable. That is true Open Science and you should be rewarded for that.

That being said, I have generated a bed file from Unibind and imported it into IGV. If I undestand correctly, all the listed TF have been SHOWN to bind direcltly  to the sequence? It is not predicted right? It is real data? There are strange codes in the bed file EXPXXX and MAXXX. Are these referring to the experiments? Where can I find them? Many thanks and my apologies if I missed something. Merci

[Anthony Mathelier]

Dear Gilles.

Thanks for your interest in UniBind and your kind words. The TFBSs are indeed of high quality as they have both experimental (ChIP-seq) and computational (JASPAR PWM) evidence for direct TF-DNA interactions. The EXPXXX codes refer to the original ChIP-seq experimental datasets while the MAXXX refer to the JASPAR TF binding profiles used.

I hope it helps.