暂态混沌神经网络在蛋白质关联图预测中应用研究

Research on Application of Transiently Chaotic Neural Network in Prediction of Protein Contact Maps

蛋白质结构预测是生物信息学的一个主要研究方向,而蛋白质关联图预测是其中的一个重要内容.针对蛋白质关联图预测问题,提出一种暂态混沌神经网络实现方法,所提出的方法具有暂态混沌特性和平稳收敛特性,能有效避免传统Hopfield神经网络极易陷入局部极值的缺陷.它通过短暂的倒分叉过程,能很快进入稳定收敛状态.仿真结果表明:暂态混沌神经网络解决蛋白质关联图预测问题时,总能收敛到全局最优或几乎接近全局最优,同时具有更高的搜索效率.这种方法预测精度达到0.27,比随机预测器高9倍.

The protein structure prediction is a main direction in bioinformaties, and the prediction of protein contact maps is an important content in protein structure prediction. A algorithm based on transiently chaotic neural network is proposed to solve the protein contact maps problem. The proposed neural networks have many merits which are transient chaos and stable con- vergence etc. so as to overcome the drawbacks of easily getting stuck in local minim in conventional Hopfield neural networks. It can reach a stable convergent state after shortly reversed bifurcations. Simulation of protein contact maps problem show the transiently chaotic neural network has higher ability to search for global optimal or near-optimal solution and higher efficiency of searching than Hopfield neural networks. The method could assign protein contacts wkh an average accuracy of 0. 27 and with an improvement over a random predictor of a factor greater than 9.