免疫-遗传算法用于混合物重叠核磁共振信号解析

通过对免疫系统中抗体对外来抗原的识别、消除等过程的模拟,建立了一种新型的免疫算法模型.将标样信号作为抗体,混合物重叠信号作为抗原输入免疫算法模型,通过迭代运算,从抗原中消除抗体所表示的信息,当抗原被抗体完全消除时,即实现了混合物重叠信号的解析.对多组分混合氨基酸NMR谱图的解析结果证明,该算法可方便地用于多组分重叠信号的解析,为利用数据库解析混合物或生物大分子等物质的复杂NMR谱图开辟了一条全新的途径.

Resolution of Multicomponent Overlapping NMR Signals Using an Immune Algorithm and Genetic Algorithm

The natural immune system has many interesting characteristics and powerful abilities from the information view points, such as recognition of antigen, production of antibody and learning with memory cell. In this paper we proposed a new algorithm by imitating the defending process of immune system to resolve the overlapping NMR signals of mixed animo acid with the aid of genetic algorithm. The method takes an overlapping NMR signal as an antigen, and the standard NMR signal of estimated components as antibodies. The antigen is subtracted iteratively by the antibodies until the residual approaches to zero and then the information of every components will be extracted. It was proven by the results that the method is a convenient tool for the resolution of multicomponent overlapping NMR signals, and is a promising way to analyze the mixtures or complex bio moleculars with the aid of database.