Experimental approaches to transient protein interactions are
laborious and serendipitous, and our understanding of fundamental
questions like the identification of interaction surfaces or the
specificity of molecular recognition between interacting proteins are
far from being solved. We propose a maximum-entropy approach based
on recent techniques from the statistical physics of disordered systems,
which exploits the natural sequence variability of homologous proteins
across hundreds of species species. Applying this approach to bacterial 
signal transduction as a test case, we show that our method is able  
(i) to identify inter-protein residue contacts and to facilitate the 
prediction of high-resolution protein complex strutures, and (ii) to 
reconstruct a protein-protein interaction code which elucidates 
specificity in signal transduction in bacteria.