基于遗传算法的神经网络为苯乙酰胺类农药构效关系建模的研究

探讨用遗传算法训练神经网络,为苯乙酰胺类化合物的QSAR建模,效果良好,神经网络可以反映复杂的构效关系,而引入遗传算法又有助于多层前传网在训练过程中跳出局部最小点,使收敛速度大大提高,并在预报精度上有显著改善.     

一种预测药物活性的神经元计算新方法

提出一种基于全变异算子遗传算法(MGA)的神经元计算新方法,用于辨识复杂药物构效关系。在MGA中,表达变量的各基因使用不同的变异概率,以便提高局部搜索效率。通过将随机初始化技术与局部搜索策略相结合,该算法能在有限时间内得到满意解。使用74个抑制还原酶的嘧啶类化合物所组成的数据集作为构效关系神经元计算的典型对象,用来考核MGA法在预测药效活性计算中的有效性。交叉验证及活性预测试验表明,用MGA法建立的构效关系模型的预测能力优于其他方法。     

Variable selection for QSAR by artificial ant colony systems

Derivation of quantitative structure-activity relationships (QSAR) usually involves computational models that relate a set of input variables describing the structural properties of the molecules for which the activity has been measured to the output variable representing activity. Many of the input variables may be correlated, and it is therefore often desirable to select an optimal subset of the input variables that results in the most predictive model. In this paper we describe an optimization technique for variable selection based on artificial ant colony systems. The algorithm is inspired by the behavior of real ants, which are able to find the shortest path between a food source and their nest using deposits of pheromone as a communication agent. The underlying basic self-organizing principle is exploited for the construction of parsimonious QSAR models based on neural networks for several classical QSAR data sets.     

Variable selection for QSAR by artificial ant colony systems

Derivation of quantitative structure-activity relationships (QSAR) usually involves computational models that relate a set of input variables describing the structural properties of the molecules for which the activity has been measured to the output variable representing activity. Many of the input variables may be correlated, and it is therefore often desirable to select an optimal subset of the input variables that results in the most predictive model. In this paper we describe an optimization technique for variable selection based on artificial ant colony systems. The algorithm is inspired by the behavior of real ants, which are able to find the shortest path between a food source and their nest using deposits of pheromone as a communication agent. The underlying basic self-organizing principle is exploited for the construction of parsimonious QSAR models based on neural networks for several classical QSAR data sets.     

Evaluation of Lipase Production by Genetic Algorithm and Particle Swarm Optimization and Their Comparative Study

This paper presents the nature-inspired genetic algorithm (GA) and particle swarm optimization (PSO) approaches for optimization of fermentation conditions of lipase production for enhanced lipase activity. The central composite non-linear regression model of lipase production served as the optimization problem for PSO and GA approaches. The overall optimized fermentation conditions obtained thereby, when verified experimentally, have brought about a significant improvement (more than 15 U/gds (gram dry substrate)) in the lipase titer value. The performance of both optimization approaches in terms of computational time and convergence rate has been compared. The results show that the PSO approach (96.18 U/gds in 46 generations) has slightly better performance and possesses better convergence and computational efficiency than the GA approach (95.34 U/gds in 337 generations). Hence, the proposed PSO approach with the minimal parameter tuning is a viable tool for optimization of fermentation conditions of enzyme production.     

A comparison of heuristic search algorithms for molecular docking

This paper describes the implementation and comparison of four heuristic search algorithms (genetic algorithm, evolutionary programming, simulated annealing and tabu search) and a random search procedure for flexible molecular docking. To our knowledge, this is the first application of the tabu search algorithm in this area. The algorithms are compared using a recently described fast molecular recognition potential function and a diverse set of five protein–ligand systems. Statistical analysis of the results indicates that overall the genetic algorithm performs best in terms of the median energy of the solutions located. However, tabu search shows a better performance in terms of locating solutions close to the crystallographic ligand conformation. These results suggest that a hybrid search algorithm may give superior results to any of the algorithms alone.     

Identifying clusters as low-lying mimina--efficiency of stochastic and genetic algorithms using inexpensive electronic structure levels

Molecular candidates possessing unconventional chemical bonding paradigms (e.g., boron wheels, molecular stars, and multicenter bonding) have attracted a great deal of attention by the computational community. The viability of such systems is necessarily assessed through the identification of the lowest lying energy forms of a given chemical composition on the potential energy surface (PES). Although dozens of search algorithms have been developed, only a few are general and simple enough to become standard everyday procedures for this purpose. The simple random search and genetic algorithm (GA) are among these: but how do these approaches perform on typical isomeric searches? The performance of three specific variants for the ab initio exploration of the PES of prototype planar tetracoordinated and hypercoordinated carbon-containing systems C₂Al₄ and CB₆²⁻ are compared. The advantages of preoptimizing with a low-cost semiempirical method (e.g., PM6) together with the most cost-efficient GA-based variant are discussed, and the trends verified by the isomer search of the larger Si₅Li₇⁺ clusters. © 2011 Wiley Periodicals, Inc. J Comput Chem, 2011     

Efficiency of random search procedures along the silicon cluster series: Si(n) (n=5-10, 15, and 20)

The efficiency of the simplest isomeric search procedure consisting in random generation of sets of atomic coordinates followed by density functional theory geometry optimization is tested on the silicon cluster series (Si(5-10, 15, 20)). Criteria such as yield, isomer distributions and recurrences are used to clearly establish the performance of the approach with respect to increasing cluster size. The elimination of unphysical candidate structures and the use of distinct box shapes and theoretical levels are also investigated. For the smaller Si(n) (n=5-10) clusters, the generation of random coordinates within a spherical box is found to offer a reasonable alternative to more complex algorithms by allowing straightforward identification of every known low-lying local minima. The simple stochastic search of larger clusters (i.e. Si(15) and Si(20)) is however complicated by the exponentially increasing number of both low- and high-lying minima leading to rather arbitrary and non-comprehensive results.     

Neural networks convergence using physicochemical data

An investigation of the neural network convergence and prediction based on three optimization algorithms, namely, Levenberg-Marquardt, conjugate gradient, and delta rule, is described. Several simulated neural networks built using the above three algorithms indicated that the Levenberg-Marquardt optimizer implemented as a back-propagation neural network converged faster than the other two algorithms and provides in most of the cases better prediction. These conclusions are based on eight physicochemical data sets, each with a significant number of compounds comparable to that usually used in the QSAR/QSPR modeling. The superiority of the Levenberg-Marquardt algorithm is revealed in terms of functional dependence of the change of the neural network weights with respect to the gradient of the error propagation as well as distribution of the weight values. The prediction of the models is assessed by the error of the validation sets not used in the training process.     

遗传算法在构象搜寻中的应用

遗传算法是一种模仿自然进化进程的新颖的启发式优化方法。通过它，人们可以在计算机上对所需的各种问题实施进化操作，最终产生理想的结果。遗传算法现已被广泛于计算机辅助设计、工程设计、系统模拟等领域并取得了极大的成功。本文拟对遗传算法在对大、中、小分子的构象搜寻中的应用作一较全面的综述。     

Implementation of the Three-Dimensional-Pattern Search Problem on Hopfield-like Neural Networks

Abstract

The three-dimensional (3D)-pattern search problem can be summarized as finding, in a molecule, the subset of atoms that have the most similar spatial arrangement as those of a given 3D pattern. For this NP-complete combinatorial optimization problem we propose, by analogy to the travelling salesman problem, a new method taking advantage of the capability of Hopfield-like neural networks to carry out combinatorial optimization of an objective function. This objective function is built from the sum of the differences of interatomic distances in the pattern and the molecule. Here we present the implementation we have found of the 3D-pattern search problem on Hopfield-like neural networks. Initial tests indicate that this approach not only successfully retrieves a given pattern, but can also suggest partial solutions having one or two atoms less than the given pattern, an interesting feature in the case of local conformational flexibility of the molecule. The distributed representation of the problem on Hopfield-like neural networks offers a good perspective for parallel implementation.     

Partial least squares modeling and genetic algorithm optimization in quantitative structure-activity relationships

Quantitative structure-activity relationship (QSAR) studies based on chemometric techniques are reviewed. Partial least squares (PLS) is introduced as a novel robust method to replace classical methods such as multiple linear regression (MLR). Advantages of PLS compared to MLR are illustrated with typical applications. Genetic algorithm (GA) is a novel optimization technique which can be used as a search engine in variable selection. A novel hybrid approach comprising GA and PLS for variable selection developed in our group (GAPLS) is described. The more advanced method for comparative molecular field analysis (CoMFA) modeling called GA-based region selection (GARGS) is described as well. Applications of GAPLS and GARGS to QSAR and 3D-QSAR problems are shown with some representative examples. GA can be hybridized with nonlinear modeling methods such as artificial neural networks (ANN) for providing useful tools in chemometric and QSAR.     

Geometry optimization in density functional methods

The geometry optimization in delocalized internal coordinates is discussed within the framework of the density functional theory program deMon. A new algorithm for the selection of primitive coordinates according to their contribution to the nonredundant coordinate space is presented. With this new selection algorithm the excessive increase in computational time and the deterioration of the performance of the geometry optimization for floppy molecules and systems with high average coordination numbers is avoided. A new step selection based on the Cartesian geometry change is introduced. It combines the trust radius and line search method. The structure of the new geometry optimizer is described. The influence of the SCF convergence criteria and the grid accuracy on the geometry optimization are discussed. A performance analysis of the new geometry optimizer using different start Hessian matrices, basis sets and grid accuracies is given.     

Ethyl alcohol production optimization by coupling genetic algorithm and multilayer perceptron neural network

In this present article, genetic algorithms and multilayer perceptron neural network (MLPNN) have been integrated in order to reduce the complexity of an optimization problem. A data-driven identification method based on MLPNN and optimal design of experiments is described in detail. The nonlinear model of an extractive ethanol process, represented by a MLPNN, is optimized using real-coded and binary-coded genetic algorithms to determine the optimal operational conditions. In order to check the validity of the computational modeling, the results were compared with the optimization of a deterministic model, whose kinetic parameters were experimentally determined as functions of the temperature.     

Partial Least Squares Modeling and Genetic Algorithm Optimization in Quantitative Structure-Activity Relationships

Abstract

Quantitative structure-activity relationship (QSAR) studies based on chemometric techniques are reviewed. Partial least squares (PLS) is introduced as a novel robust method to replace classical methods such as multiple linear regression (MLR). Advantages of PLS compared to MLR are illustrated with typical applications. Genetic algorithm (GA) is a novel optimization technique which can be used as a search engine in variable selection. A novel hybrid approach comprising GA and PLS for variable selection developed in our group (GAPLS) is described. The more advanced method for comparative molecular field analysis (CoMFA) modeling called GA-based region selection (GARGS) is described as well. Applications of GAPLS and GARGS to QSAR and 3D-QSAR problems are shown with some representative examples. GA can be hybridized with nonlinear modeling methods such as artificial neural networks (ANN) for providing useful tools in chemometric and QSAR.