A Correction Algorithm for the Torque/Rotation Coefficient Used in the Prediction of Protein Conformations Using Robotic Methods

Abstract

The requirement to study and simulate the behavior of biomolecules such as proteins is increasing. Consequently, the introduction of new methods that can predict the conformation of proteins faster and more accurately is important. A recently proposed method in this field is the PROTOFOLD algorithm. Robotic techniques such as the zero reference position notation method and the Amber force field, in addition to a mechanical model for chemical bonds with a constant coefficient, are used to model protein structures. In this report, an algorithm is presented that optimized and decreased the solution time using the PROTOFOLD approach. Since a constant torque to rotation coefficient can cause problems in solution, the presented algorithm allows changes to the coefficient in response to the solution conditions, and thereby reduces the time required for the modeling. These solution conditions include an increasing energy trend, the stability of the molecule, or torques values applied on bonds in comparison with other bonds and other steps. So the step resolution of the solution reduces and the protein reaches its stable configuration faster. The results are presented for the protein 4rxn. Here, the protein normally reaches its stable form in 100 steps; however, using the proposed algorithm a stable conformation is reached in 60 steps.