Section Article

A Refinement of the Particle Swarm Optimisation Model for Predicting Macromolecular Structures
Author(s): Jaskaran Singh

Abstract
Macromolecular structure prediction has emerged as one of the most computationally complex and biologically significant domains in modern computational biology structural bioinformatics molecular modelling and protein engineering. Accurate prediction of three-dimensional macromolecular conformation particularly proteins nucleic acids and large biological assemblies directly influences fundamental scientific tasks such as drug design enzyme engineering protein–ligand docking identification of functional motifs and simulation of biological pathways. Over the last two decades metaheuristic algorithms have been increasingly applied to the high-dimensional search spaces associated with molecular structure prediction. Among these Particle Swarm Optimisation has become one of the most extensively studied nature-inspired algorithms because of its simple implementation collective intelligence modelling reduced computational cost and strong capability to explore complex multidimensional landscapes. However classical PSO suffers from several structural limitations when applied to biological molecules: premature convergence stagnation in local minima instability under rugged energy landscapes weak adaptability to macromolecular folding patterns difficulty in maintaining population diversity and limited capacity to integrate physico-chemical constraints. These limitations restrict PSO’s capability to accurately approximate the global minima of macromolecular potential-energy surfaces.