Centre de Bioinformatique, INSERM U155 Universite Paris 7, tour 53, 1er etage 2, place Jussieu, 75251 Paris, France
Despite notorious efforts, the techniques of simulation remain unable in general to explore the ways of the folding, allowing the protein to take its native conformation. The aim of this study is to predict a set of "candidate contact points" from the knowledge of the sequence. From this set , it will be possible to tackle, in a further study, the problem of the building structural models of a given protein.
A previous statistical study [1] allowed the extraction of the
preferential interactions between the side-chains, first independently of
their locations in the 3D structure, and secondly, taking the accessibility
to the solvent of the two concerned side-chains and their locations in
secondary structures into account. The main result of this study is that the
association residue-burial contributes on the interactions more highly than
the association residue-secondary structure or residue type-residue type.
From this analysis, a potential of mean force was defined taking the
side-chain interactions, the residue burial and the location in secondary
structure into account. The aim of this study is to focus a method of search
for possible contacts in a given protein from the knowledge of the primary
sequence and this potential of mean force.
We have started to explore whether it is possible, from the knowledge of the primary sequence, the secondary structures and the exposition to the solvent, to find again the contacts observed in proteins of known 3D structure, in defining a "dot-plot" allowing the prediction of "candidate contact points". A first method is an assessment of the contact prediction deduced by a simple thresholding of the potential. A possible second method is the using of a learning system such as a neural network. The surroundings of residues will be taken into account in the study and will used either directly in the potential or can be taken in the learning into account. Perspectives of such strategies when residue burial and/or residue secondary structure are taken into account, are discussed.
From the knowledge of the "candidate contact points", a further work will consist in simulating a set of tridimensional protein models in order to try to find again the real protein structure.
[1] Mucchielli-Giorgi, M.H., Tuffery, P. and Hazout, S., (1996). Statistical assessment of the major contributions driving side-chain interactions. Submitted.