Find The List Of Tf Likely To Bind To A Promoter Or A Genome Location

Question

[Sirus]

Hello everybody, I have list of genomic positions (promoters) and I want to know what are the TFs that are likely to bind there. I think there should be some methods developed to assess if a TF can or cannot bind to that location given for example the sequence.

How do you guys generally do that, is it just take the motif of TF and align to the region of interest and then check how good it aligns there, and then take the TF that best align to that place ? or there are more sophisticated methods?

Thanks in advance :)

Answer 1

[KCC]

I am not sure about the best way to do what you ask for, but I thought I would take a stab at your question. I think the extent to which you can answer this question will be organism and TF specific. Some TFs have well-defined motifs associated with them. If the TFs you are interested in are some of those, then you are part of the way toward your answer. Your ability to answer your question also depends on how much information has been gathered about your particular TFs. Have they been studied before? Is the sequence they bind to known? If you have information from a protein-binding microarrays for instance, you can make some guesses as to the promoters that your TF is likely to bind to. In addition, you would need some idea of what parts of your promoter are accessible to filter out false positives. If your TF can't access a certain part of the genome, then it doesn't matter if it could theoretically bind strongly to that area.

Another approach is to have data on what parts of the genome each TF binds to from ChIP-seq or something similar. This data would tell you directly whether your TF has binding sites in the promoters of interest. Then you can do simple overlaps.

Accessibility and binding sites are life-stage and tissue specific. DNA-seq data gives you an averaged picture over tissue types and whatever stages are in your data. So, it's certainly not perfect.

In most cases, this question you are interested in is not a simple project. It can take years of wet lab work and bioinformatics depending on the level of accuracy you are seeking.

[Sirus]

Thanks George for your very clear answer and good advice. Yeah, indeed it is not a trivial question to answer, because as you said, it is tissue and life-stage dependent. I will try check if other people have tried to develop some heuristic methods for that other other wise I will find my self obliged to develop one :(.

Answer 2

[Alex Reynolds]

Take a look at FIMO ("Find Individual Motif Occurrences"). You input your promoter sequences and motif information (JASPAR, etc.). It tells you what the likelihood is of finding a particular motif in your sequence regions.

Answer 3

[Ming Tommy Tang]

Try MEME http://meme.nbcr.net/meme/