| Evolutionary algorithms are a class of random search algorithm, which reference to natural selection and natural genetic mechanism of biosphere. Evolutionary algorithms have the following main aspects:Genetic Algorithm (GA),Genetic Programming (GP),Evolutionary Strategies (ES),Evolutionary Programming (EP),Particle Swarm Optimization (PSO) algorithm and Quantum-behaved Particle Swarm Optimization(QPSO) algorithm which has developed at present. They are capable of finding the Optimum of the problem through a set of evolutionary operators and evolutionary equations. This paper solve many problems of bioinformation by evolutionary algorithms and their improved algorithms, coupling mathematical models。The main contents are outlined as follows.(1) Due to shortcoming of Quantum-behaved particle swarm optimization (QPSO) that it is often premature convergence, two improved QPSO algorithms are presented by integrating selection mechanism. The selection operator is exerted on the global best position to improve the search ability of the QPSO algorithm. One is the QPSO with tournament selection (QPSO-TS) and the other is the QPSO with roulette-wheel selection (QPSO-RS). Furthermore, we prove the algorithm is global convergence from theory. In the experiments, twenty functions were used to validate the performance including benchmark Unimodal and Multimodal functions, Rotated Unimodal and Multimodal functions and shifted Unimodal and Multimodal functions. The experimental results show that the improved QPSO algorithms have powerful optimizing ability and higher optimizing precision.(2) Multiple sequence alignment (MSA) is a fundamental and challenging problem in the analysis of biologic sequence. In this paper, we propose two methods to solve MSA. One method is mutation-based binary PSO (MBPSO) and binary QPSO (MBQPSO) for MSA solving. In the proposed MBPSO and MBQPSO algorithms, BPSO and BQPSO algorithms are conducted to provide alignments. Thereafter, mutation operator is performed to move out of local optima and speed up convergence. Another method combined QPSO algorithm and improved QPSO algorithms with Profile Hidden Markov Model (HMM) for MSA. From simulation results of nucleic acid and amino acid sequences, it is shown that the proposed algorithms can find out optimal alignments of multiple sequences.(3) Metabolic flux estimation through 13C trace experiment is crucial for metabolic system to quantify the intracellular metabolic fluxes. In essence, it corresponds to a constrained optimization problem, objective function of which is non-linear and non-differentiable and exist multiple local minima making this problem a special difficulty. In this paper, two methods were proposed to solve 13C-based metabolic flux estimation problem. One method is a self-adaptive evolutionary algorithm with singular value decomposition. Another method is Quantum-behaved particle swarm optimization (QPSO) and improved QPSO algorithms with penalty function. The proposed algorithms are applied to estimate the central metabolic fluxes of Corynebacterium glutamicum and compared with conventional optimization technique. Experimental results illustrated that our algorithms are capable of achieving fast convergence to good near-optima.(4) For optimizing the culture conditions (agitation speed, aeration rate and stirrer number) of hyaluronic acid production by fermentation. GP is employed to model the microbial HA production and QPSO algorithm is used to find out the optimal culture conditions with the established GP estimator as the objective function. The experimental results indicate that GP-QPSO could produce the maximal production of hyaluronic acid. Here though both RSM and GP-QPSO approach provided good predictions, yet the proposed GP-QPSO method showed a clear superiority over RSM for both data fitting and optimization capabilities. With the advantages of rapid convergence and global optimization, the proposed GP-QPSO approach is expected to be widely applied in bioprocesses optimization.
|
PDF Full Text Request