基于多级滤波器的蛋白质热点区域预测

摘要

针对数字滤波技术用于蛋白质热点区域预测时预测结果精度低的问题,提出一种基于多级滤波器的预测方法.依据共振识别模型从一组蛋白质中提取出特征频率,设计一个选通特性良好的多级滤波器对蛋白质序列进行滤波,根据滤波器输出序列的能量谱定位蛋白质的热点区域.实验表明,多级滤波器能提高预测结果的精度.%Aiming at the low accuracy of digital filter to identify the locations of hot spots in proteins, this paper presents a new technique of using multi-stage filter. The technique selects characteristic frequency from a set of proteins based on resonant recognition model, and designs a special multi-stage filter to process the protein sequence, and then the energy spectrum of the filter output yields the locations of the hot spots in proteins.Experiments show that this filter has higher accuracy.