An Object-Oriented Architecture of the Information System for Solving the Protein Folding Problem

The paper presents an architecture of the information system designed for dealing with the problem of the protein tertiary structure's prediction. The system is implemented through the object-oriented approach. The critical properties of the given process, such as abstraction and polymorphism, make the developed architecture open and extendable. The system's concept includes two abstract concepts, the “Problem” and the “Algorithm.” Within the framework of both concepts, some interfaces have been implemented that determine the system's structure, interconnections, and behavior of its particular components. A specific implementation of the features in the “Algorithm” part includes such algorithms as clonal selection, differential evolution, and a hybrid version of the mentioned procedures. The “Problem” is represented by models of the protein folding on two-dimensional and three-dimensional square lattices as well as a model using torsion angles for a description of a molecule's shape, for which energy is calculated employing a potential energy function that takes into account the molecular dynamics of a protein. The system's core is implemented in C++. The architecture includes several interfaces which make it possible to use the system as a Python extension.