Protein Structure Predicted By Swarm Intelligence

Protein structure prediction is the prediction of the three-dimensional structure of a protein from its amino acid sequence-that is,the prediction attempts to determine the native structure of its given primary structure.Protein structure prediction is one of the most important goals pursued by bioinformatics and theoretical chemistry;it is highly important in medicine and biotechnology,such as drug design,and the design of novel enzymes.As an important term in computational prediction,knowledge of protein structure determinants is critical:the hydrophobicity and hydrophilicity of residues,electrostatic interactions,hy drogen and covalent bonds,van der Waals interactions,bond angle stresses,and enthalpy and entropy.Another important factor is to develop an efficient computational optimization.In this study,several computational intelligence methods are developed for individualized models and protein structure predictions.The study involves:(1)Chapter 2 presents three main search algorithms including heterogeneous particle swarm optimizer,nonhomogeneous cuckoo search algorithm and nonhomogeneous firefly algorithm.In the heterogeneous particle swarm optimizer,the population is grouped into four sub-swarms to maintain heterogeneous search strategies.The information sharing mechanism not only gathers the useful messages from each sub-swarm,but also contributes to the cooperation and potential search abilities.In the nonhomogeneous cuckoo search algorithm,the individuals can learn the difference between pairs of agents and also they can potentially learn from the averaged information of the whole swarm,as well as the global best information based on the quantum mechanism for efficiently search.The novel nonhomogeneous update laws make the individuals have efficient search abilities in both local regions and large potential space.In the nonhomogeneous firefly algorithm,the values of a in the five different strategies decrease dynamically with the generation number,population size and the size of the optimized problem,while in the algorithm a scheme is proposed to smooth balance the exploitation and exploration in the search process.A distance-based technique is developed to overcome the two main drawbacks in using a constant light absorption coefficient,which tunes the light among the fireflies dynamically to control the sharing distance information leading to the variation of the attractiveness.Simultaneously,the differences among the fireflies can be adequately used to enhance the local search ability of each firefly,hence we employ the gray relational analysis to design a gray coefficient as another self-adaptively altering parameter.We have also theoretically proved that the three algorithms can converge.The analyses provide theory for the applications of individualized models and protein folding pathways and structure predictions.(2)Chapter 3 presents individualized models based on fuzzy system and neural network for biological sy stems and protein fold recognition.In fuzzy system,the biological systems can be modeled by a rule-based approach with the optimal structure and the associated suitable parameters.Moreover,it allows to efficiently address questions on the dynamic behaviors of the systems biology robustly with transparent and understandable rules.To recognize protein folds,an individualized model is developed by encoding several potential neural networks into different particles of an enhanced heterogeneous particle swarm optimizer.and combining evolutionary-and structure-based information of amino acid sequences,the model can recognize the protein folds with the optimal parameters and structure of neural network.(3)The goal of protein folding pathways and structure prediction is to find out how the protein folds itself and what the spatial position of every atom is from the amino acid sequence by computational methods.Chapter 4 presents t,he computational methods including opt,imization methods and molecular dynamics,for protein folding pathways and structure prediction.Based on potentials of physical and chemical properties,we computationally infer protein folding pathways and structure from its amino acid.