(a) Which is the best database for transcription factor binding site, TRANSFAC or JASPAR? Why?
(b) Is it still valid to assume, "transcription factor binding sites are only present in 5'UTR and in promoter region"?
Note: kindly put references in your answer.
This very much depends on what you are trying to accomplish.
TRANSFAC and JASPAR are databases of transcription factor binding models. Using experimental techniques like SELEX and now increasingly ChIP-Seq we can identify the kinds of sequences that a particular transcription factor (TF) has affinity for. This affinity is typically represented with something like a position weight matrix (PWM) or position-specific scoring matrix (PSSM). Those models can be used to scan a genome for "putative TF binding sites" that have a high similarity to the model. Whether those sites are actually bound by a TF and play a functional role in transcription and under what conditions must still be determined experimentally be traditional molecular techniques like promoter bashing, reporter gene assays, ChIP experiments, etc.
Databases like ORegAnno and PAZAR have attempted to collect and curate individual experimentally proven TFBSs from the literature and sets from ChIP-Seq and other genome-wide experiments. Technologies like ChIP-Seq can give us a genome-wide profile of all the binding sites for a particular transcription factor under a specific condition. Many of these TF binding sites (TFBSs) are being deposited into species-specific databases (FlyBase, SGD, etc) and as tracks in the UCSC browser. In particular, see efforts by ENCODE. A good starting point would be to look through the tracks under 'Regulation' in the UCSC Genome Browser for your species of interest.
I just happened to ask how JASPAR compares to Transfac to members of the JASPAR team present at a meeting in Montpellier, I think about 2 years ago. The answer I got was "Transfac (the Pro version) covers more, but our quality is better".
If you mention Transfac Pro I think you should also mention Genomatix. They stem from the same basic Transfac development in Germany. There were two commercial lines started from this project and they both developed upon the original open Transfac. They were in fact early endeavors to see whether government funded investments in databases could be made sustainable through offering commercial services.
Academic users can use most of parts of the Genomatix toolset and database with a free account. Access to (part of?) the commercial content of the Transfac Pro is possible through websites that use it behind the scenes.
You might also want to consider the UniPROBE database which contains PSSM models based on protein-binding microarrays (PBMs).
If you are doing your analysis in yeast, you should consider the YeTFaSCo database which is a curated collection of PSSM models for most of the yeast TFs.
Hundreds of transcription factor (TF) binding preferences have already been queried using PBMs, soon thousands will be. JASPAR and TRANSFAC are a bit slow in keeping up with these new data, it may be worth it to keep an eye on the work of Tim Hughes and Martha Bulyk who are the ones actively generating TF binding preference data using PBMs.
Just to show that the free version 3.2 of transfac is really old: Version 3.2 is from 1997.
For PSSMs/PWMs of human/mouse TFs you may also try HOCOMOCO (http://hocomoco.autosome.ru).
You can access the TF binding sites in Plants at PlantRegMap, which are derived from ChIP-seq and DNase I genomic footprinting.
Genexplain recently uploaded a white paper explaining the difference between TRANSFAC and JASPAR. It seems when we use TRANSFAC there are chances that we don't miss important matrices that play a role in regulation of gene of interest. The white paper link can be found here:
Hopefully this is still helpful to someone.
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
I think it is pretty optimistic to think that there is a reference to document that one specific database of TF binding sites is better than another. How would you even define "best" in this case?
Changed title to be more descriptive
True, I agree, but what should be the preference while selecting one database and not the other.