Protocol Online logo
Top : New Forum Archives (2009-): : Bioinformatics and Biostatistics

software to predict domains to proteins interactions - (Oct/31/2013 )

Hi there¡¡

Im studying a protein involve in transcription activation in yeast. The cristalographic structure of it human version was solved by others authors by X-ray and NMR. Now, we are research the functional division of its differents domains and we need a first aproach to start to design new experiments. 

We have a set of factors (aproximately 20) that could be associated with our protein. We have confirmed that one of those factors is attached to our protein, but we do not know to which domain does it.

I would like to known if ithere is a software to help us to find putative regions inside of structure that could interact with others protein (including itself).

Thanks in advance.




There are a few protein-protein interaction predictors out there, just google for them.


You will still have to experimentally verify this interaction - the easiest (and classic) way of doing this is to create a series of truncation mutants and see which ones bind and which don't, then further refine based on the presence of binding.


Thanks bob.

But I would like to know the putative region inside of the secuence/structure of my protein used to interatc with other protein?.is it possible to predict it using a bioinformatic software?

Thanks again.


It is only really possible for some of the many interactions that are known - not many of the interactions seem to be through a consistent sequence, but for those few, and based on biochemical similarities, it can be possible to identify the sites.


Yes, thanks for you replay.

I have followed your advice, and recently, we confirm a direct interaction between "my" protein and one of the candidates, whose crystallographic structure was solved 4 years ago. (Of course, the sequence of both yeast proteins is known).

Now, we are interested in determine which domain/module/zone of my protein is involve in such interaction, of course, one alternative could be with a set of delectants mutants, but I want to reduce the spectrum of possibilities with a prediction by software.That information will determine the approach and future experiments of my thesis. So, is there a software able to predict the domain used for interact with other proteins?

Thanks again¡¡¡



Short answer: no.


Long answer: maybe - like I said before, there are many many different protein:protein interactions, some known, some unknown.  Few interactions between different proteins seem to be through consistent domains, but there are exceptions (e.g. poly-proline regions and SH3 domains, LXCXE and pocket proteins).  All you can do is look around for protein:protein interaction software, and see what domains it pulls out.