Skip to main content

COMPUTER-AIDED PEPTIDE AND PROTEIN ENGINEERING SOFTWARE DEVELOPMENT

Objective

THE PRIMARY AIM OF THE SYSTEM WILL BE AS A TOOL FOR PRECALCULATION THE EFFECTS OF SUBSTITUTING NEW AMINO ACIDS IN PROTEINS BY SITE-DIRECTED MUTAGENESIS, I.E. THE DESIGN OF NOVEL PROTEINS DERIVED FROM EXISTING ONES OF KNOWN CONFORMATION. IN ADDITION, THE SYSTEM WILL RETAIN THE CHARACTER OF CURRENT SOFTWARE IN BEING APPLICABLE TO ALL CONFORMATIONALLY FLEXIBLE MOLECULES, AND TO THE PREDICTION OF PROTEIN CONFORMATION FOR AMINO SEQUENCE ALONE AND HONOLOGOUS PROTEINS.

IT SHOULD BENEFIT ACADEMIC AND INDUSTRIAL RESEARCH LABORATORIES BY SPEEDING UP THE DESIGN PROCESS OF NOVEL PROTEINS.
For a rational design of new recombinant proteins one needs to understand the relationship between the chemical structure (amino acid sequence) and the function of the protein. The insufficient knowledge of these relationships represents the real bottleneck in the development of newly engineered proteins or synthetic vaccines. However, new understanding gained with computers provides a possibility to improve protein modelling and design.

A program, Pred 85, which included a rule based approach and secondary structure prediction, was tested for the prediction of external regions of proteins to selecting peptides for vaccines. In parallel, new secondary structure prediction methods have been improved and developed to incorporate into Pred 85 and for the modelling of various proteins. With the guidance of the University of Manchester Graphics Unit new graphic facilities have been developed for protein molecular modelling.
The project was successful in:
producing methods based on amino acid sequence alone for predicting secondary structure with an accuracy of 65.5 to 87%, whereas previous methods were in the range of 49 to 55%;
producing a graphics systems of high quality for modelling complex protein structures;
testing a programme based on sequence alone to predict epitopes (a raised antibodies against 2 proteins (hormones) with synthetic peptides one of which would not have been detected by conventional prediction techniques);
modelling and detailed analysis of protein sequences and discovery of their relations to other proteins by homology or by 'expert system' rules (demonstrated by a widely acclaimed application to epidermal growth factor receptor);
developing accurate and efficient sampling of thermodynamic properties by hybrid dynamics or simulated annealing.
THE PROPOSED PROJECT WILL INVOLVE COMBINING PRESENT AND FURTHER SOFTWARE WITH INTERACTIVE MOLECULAR GRAPHICS. THE SYSTEM WILL BE CONTROLLED BY A USER-FRIENDLY LANGUAGE-LIKE COMMAND STRUCTURE. IT IS INTENDED TO DEVELOP A NEW SOFTWARE BASED ON INFORMATION THEORY AND PATTERN RECOGNITION ALGORITHMS TO DETECT THE MOST PROTEIN ANTIGEN (I.E. THE PARTS OF THE AMINO ACID SEQUENCE WHICH ARE RECOGNIZED BY AN ANTIBODY). IN THE CASE OF ALREADY SUCCESSFUL SYNTHETIC PEPTIDES COMPUTING PROGRAMMES WITH INTERACTIVE MOLECULAR GRAPHICS WILL BE DEVELOPED TO TAILOR MORE STABLE NATIVE CONFORMATIONS BY SEQUENCE MODIFICATIONS YET RETAINING THERE IMMUNOGENICITY I.E. THIS ABILITY TO ELICIT AN IMMUNE RESPONSE

IN RELATION TO THIS, ATTENTION WILL BE GIVEN TO A NUMBER OF INTERESTING SYSTEMS (E.G. GROWTH HORMONE, CHORIONIC GONADOTROPIN, EPIDERMAL GROWTH FACTOR AND GROWTH HORMONE RELEASING FACTOR) AND IT IS LIKELY THAT THOSE FOUND PARTICULARLY TRACTABLE AND OF EMERGING MOST GENERAL INTEREST WILL BE GIVEN SPECIAL EMPHASIS. VALIDATION OF PREDICTED ANTIGENICITY WILL BE PERFORMED WITH SYNTHETICALLY DERIVED PEPTIDES.

Funding Scheme

CSC - Cost-sharing contracts

Coordinator

NOPA / NOeMI Research Unit
Address
Inra, Blg 440
78352 Jouy-en-josas
France

Participants (2)

University of Manchester
United Kingdom
Address
Oxford Road
M13 9PL Manchester
Wellcome Research Laboratories
United Kingdom
Address

BR3 3BS Beckenham