The preferred binding site for PPARg is known; the DR1 (AGGTCAnAGGTCA). For less well characterized TFs, you need to look at high-throughput screens that look at motif preferences via SELEX or protein binding microarrays. For examples, see the following papers:

"Determination and inference of eukaryotic transcription factor sequence specificity." Cell. 2014 Sep 11;158(6):1431-43

"DNA-binding specificities of human transcription factors." Cell. 2013 Jan 17;152(1-2):327-39

Even these high-throughput assays will still leave some targets orphaned in terms of finding a preferred motif for a TF of interest, but they are the best there is to offer right now.

Hi Aaron

this is easy for you

you can align your DNA sequence and Transcription factor sequence by using BLAST

*I have modified my original question to make it a bit more clear*, but what I was originally trying to express is that I do not know the identity of the sequence that the transcription factor binds to. That is what I am looking to find at this point.

Thanks for the reply, Vinod- I agree that the last step should be very easy to align once I can figure out the consensus sequences. I appreciate the response.

The preferred binding site for PPARg is known; the DR1 (AGGTCAnAGGTCA). For less well characterized TFs, you need to look at high-throughput screens that look at motif preferences via SELEX or protein binding microarrays. For examples, see the following papers:

"Determination and inference of eukaryotic transcription factor sequence specificity." Cell. 2014 Sep 11;158(6):1431-43

"DNA-binding specificities of human transcription factors." Cell. 2013 Jan 17;152(1-2):327-39

Even these high-throughput assays will still leave some targets orphaned in terms of finding a preferred motif for a TF of interest, but they are the best there is to offer right now.

If you are playing with a transcription factor that's is well study (known binding consensus sequence), transfac or tfsearch will be a good site for you.

If you are dealing with a novel transcription factor, you can either go for the bioinformatics approach which to predict the DNA binding domain of your factor and then its potential consensus sequence (will be a long shot and require massive downstream verification). The other easier approach will be doing a ChIP-seq (if there is specific antibody available for your factor, or you can raise your own which should be worth the effort).

Just my two cent, hope this will be helpful for you.

Once you find out the consensus sequences and have a method to apply this bioinformatically - be cautious in the interpretation. Remember that in silico motif finders presuppose that all DNA is accessible to the transcription factor, when clearly it is not. Nucleosomes, epigenetic regulation of the DNA and so on will inhibit binding. The transcription factor may also need a cis enhancer element to bind adequately.

As a test, find whatever relevant empirical data you can and see how it matches your in silico predictions.

Thanks for the input, everyone! 

I ended up finding some putative consensus sequences, by looking at recent publications that analyzed ChIP-seq data (which i hadn't found before!). I wasn't able to find any info about the less-well known TF's in the public TRANSFAC database.

Then I had issues finding a program that could align the consensus sequences using the DNA ambiguity codes (to account for the relative confidence of each nucleotide in the consensus sequence). But i managed to find a site that worked very well for this analysis, called Yeastract: http://yeastract.com/formtfsbindingsites.php 

And i found a possible match in my promoter sequence! Yay! But I will heed your words of caution, Jason.

http://yeastract.com/formtfsbindingsites.php

A site that I've found quite useful is LASAGNA-Search 2.0 at:

http://biogrid-head.engr.uconn.edu/lasagna_search/

You can select TF motifs from a variety of different databases, choosing to screen only the ones of your interest or even define your own to search your sequence.  For each motif, the site provides the WebLogo image for the factor as well as links to references.  Much more flexible than TFSearch, which has also actually been inaccessible/non-functional for many months.

Well, first of all, consensus sequences are not a good idea for prediction of TFBSs. For known sites of known TFs there are databases mentioned above (I would also suggest Jaspar and its searching tool, which gives less false positives than others BUT digests only 1 promoter per run, so if you work with a particular promoter, it's a good choice.

In case you have a set of known/predicted binding sites for not well known TF (not described in databases so far) and you want to predict the binding sites for this TF in your sequence, I would recommend SiTaR (https://sbi.hki-jena.de/sitar/) or if you're interested I can send the standalone version. For SiTaR, you don't need any consensus or matrix, you just use the sites/motifs as they are.

And remember that all these binding site predictions are very much error-prone! The number of false positives is usually high but with some experience of playing with cut-offs you will get nice results:)

Hope this helps

I agree with Shelest that consensus sequences are not a good idea to determine TFBSs, and furthermore, ChIP-seq basically provides opportunity to predict targets of your TF, but needs a lot of expense to do that. If you already have the prediction about the target then it is simply better to avoid ChIP-seq. It is better to map your primers for qPCR to validate the target of the TF followed by performing EMSA. Map your target at the up-stream or down-stream to the TF start side. In a subsequent step, perform EMSA using your appropriate primer pairs (target size might be 100 to 200 bp) followed by DNA foot-printing as to get potential TF binding site in the promoter.   

 Thank you all for the very valuable input. i will be sure to consider all of these suggestions as i continue forward.

Aaron, you could use our MatBase database and MatInspector software to get a list of weight matrices for TFs (MatBase) and search them in arbitrary DNA sequences (MatInspector). MatInspector also will allow you to search for IUPAC consensus strings. Also if you have a few sites, you can generate your own consensus sequence / weight matrix using CoreSearch and use this for a MatInspector search. A trial version of the software is available at the link attached. I've also added a publication for the software. Let me know if you have any questions.

http://www.genomatix.de/cgi-bin/UMapps/register.pl

Article MatInspector and beyond: Promoter analysis based on transcri...