I used the PROMO site and found it to be very accurate. I performed several functional assays to corroborate the predictions, finally performing SILAC mass spec on the purified TFs to confirm identities of TFs binding to different alleles of particular SNPs. For my sequences, PROMO was the most accurate, although I did initially use many other sites, and I liked it because it was easy to use, was easy to see the binding sites in relation to the sequence as they are shown in different colours and is species-specific.

The PROMO site combines information from different sites, including TRANSFAC.

(http://alggen.lsi.upc.es/cgi-bin/promo_v3/promo/promoinit.cgi?dirDB=TF_8.3)

I don't think there is online software for this, since it is a complex calculation to perform. You need a very powerful computer that can perform docking (http://en.wikipedia.org/wiki/Docking_%28molecular%29).

http://www.cbrc.jp/research/db/TFSEARCH.html

http://alggen.lsi.upc.es/cgi-bin/promo_v3/promo/promoinit.cgi?dirDB=TF_8.3

It can identify putative transcription factor binding sites in DNA sequences from a species or groups of species of interest. The latter is useful if you suspect that the TF site is conserved across species.

Http://zlab.bu.edu/zlab/gene.shtml

http://www.sabiosciences.com/chipqpcrsearch.php?gene=GADD45A&factor=Over+200+TF&species_id=0&ninfo=n&ngene=n&nfactor=y

I like the rsat (http://rsat.ulb.ac.be/) set of tools for cis-element analysis.

I have also found POCO and Softberry useful at times:

http://ekhidna.biocenter.helsinki.fi/poxo/poco/

www.softberry.com

try DiRE for evolutionary conserved sequences

For transcription factor binding site prediction you can use these tools

TRANSFEC (http://www.gene-regulation.com)- from Biobase - Not freely available

JASPER ((http://jaspar.genereg.net) - Open Access

Information of many other online tools is available on following link-

http://molbiol-tools.ca/Transcriptional_factors.htm

Hi, the ability of the TF to bind or not a DNA sequence depends mainly on its structure. You first need to model the structure of the TF in order to infer the binding site. There are a couple of servers that will do the work for you based on statistical potentials:

http://www.gene-regulation.com/pub/programs/3dtf/

http://dna.bime.ntu.edu.tw/pidna

I used the PROMO site and found it to be very accurate. I performed several functional assays to corroborate the predictions, finally performing SILAC mass spec on the purified TFs to confirm identities of TFs binding to different alleles of particular SNPs. For my sequences, PROMO was the most accurate, although I did initially use many other sites, and I liked it because it was easy to use, was easy to see the binding sites in relation to the sequence as they are shown in different colours and is species-specific.

The PROMO site combines information from different sites, including TRANSFAC.

(http://alggen.lsi.upc.es/cgi-bin/promo_v3/promo/promoinit.cgi?dirDB=TF_8.3)

Find two useful website:

http://mirstart.mbc.nctu.edu.tw/browse.php

http://www.genetools.us/genomics/Promoter%20databases%20and%20prediction%20tools.htm

We have recently a new software pipeline called Haystack find TF enriched :

https://github.com/lucapinello/Haystack/blob/master/README.md

It also integrate gene expression data and epigenetic data if you have.

We validated the pipeline in this PNAS paper:

http://www.pnas.org/content/111/3/E344.abstract

Any feedback is well appreciated!

How to interpret the result of PROMO software.

 has anyone used CONSITE to predict trnscription factor binding sites?

 is it reliable? and what score cut off is generally takes so as to avoid false positive hits predicted by CONSITE?

You can use this tool to preditc TFBS and CRMs as well. It allows to predict on a genomics scale and add rules to limit the rate of false positives.

http://bioinformatics.ibioba-mpsp-conicet.gov.ar/INSECT2/

Does anyone know a tool that would show CREB half-sites (TGACG, CGTCA)? PROMO/Jaspar seem to ignore them.

Misha, in a PWM which is the type of matrices used by these type of programs, positions are assumed independent. You can just cut the matrix into two, and use each half separately to look for your sites.

You can use this tool to predict sites.

http://bioinformatics.ibioba-mpsp-conicet.gov.ar/INSECT2/

Misha, CREB binds DNA as an homo- or heterodimer. This is the reason why JASPAR/PROMO, which contain experimentally determined TF-motif profiles, cannot detect CREB half-sites, because they contain the dimer profile. Your best shot would be splitting the JASPAR/PROMO CREB profile in half and scan your sequence using FIMO, for example.

Hi Misha, I just pasted as FASTA sequence into ConSite, selected CREB from the available TF choices, clicked "analyze", then selected "sequence view " from the left hand side menu option.  This shows my sequence with all the CREB binding sites. Some of them contain the half-sequences you were looking for. Depending on the length of your sequence this may be one way to check your sequence fairly quickly.

Hi Misha,

How was your experience working with this? As far as I am concerned, Jaspar is the most informative tool to work with. However, it is not very intuitive. I am trying to download all of the consensus sequences of Transcription factor response elements. But no luck so far! Did anyone here have an idea how to do it? Thanks

Hi Trung, we are currently working on the next JASPAR release. If you contact me, I can send you the consensus binding sequences by e-mail.

Katherine, ConSite seems to work, thanks! Looks like this tool could be used when one doesn't know the responsible TF binding sites beforehand and wants to find potential candidates.

Although, it doesn't show degenerate SRF binding sites (which are functional in some cases).

I´ve tried lots of TFBS web predictors and I´m astonished with the discrepancy between all of them... I can not believe there isn´t a decent concensus regarding this.

Maxi Ininfa in general, TFBS predictions depend on 1) a profile (i.e. PFM, PWM, TFFM...) and 2) a threshold (only predictions above that threshold are kept). The better the profile and the more stringent the threshold, the more reliable the predictions will be. Note that the quality of the profile depends not only on the computational model (e.g. TFFM) but also on the experimental data used to derive the model (e.g. ChIP-seq, HT-SELEX, PBM...). Moreover, intersecting predictions with ChIP- and/or DNase-seq data will result in even more reliable predictions. For a direct map of TF-DNA interactions you can refer to MANTA2 (https://www.nature.com/articles/sdata2018141) or UniBind (http://unibind.uio.no). TFBSs were predicted using JASPAR profiles within ReMap ChIP-seq datasets for that TF. Note that while these TFBSs are reliable, they don't include ALL TFBSs.

Katherine Thain

Hi, I am trying to use PROMO to predict transcription factor binding sites for the enhancer of interest. You seem pretty experienced with PROMO website. Do you know what the color coding for the predicted transcription factors means? Also the numbers in the colored box for each transcription factor? Thank you.