Application Of Membrane Computing In The Analysis Of Gene Expression Data

Membrane computing is a new branch of natural computation; it’s a novel class of computing models inspired by the structure and functioning of living cells as well as from the cooperation between cells in tissues and organs. This computing model called P system generally. Because of its distributed and parallel computing capabilities and the advantages and features that different models have shown, membrane computing has attracted the attention of many scholars, and it was widely used in many fields.With the completion of multiple biological genome sequencing and the wide application of microarray technology, the explosive growth of gene expression data urgent needs automatic and effective data analysis tool. Thus, the analysis of gene expression data has become a hot and difficult research in Bioinformatics. Currently, the clustering analysis has become a powerful tool for the analysis of gene expression data in Bioinformatics.At first, this thesis combines the fuzzy partition technology and mechanism of genetic evolution, and using real number coding method to encode the initial objet, thus can to obtain a good clustering effect on low dimension data set. Based on this, because gene expression data has many characteristics such as many types, large capacity; inner relationship between the data is more integrated and complex, etc. To analyze the gene expression data, this thesis presents a fuzzy clustering algorithm under the membrane computing framework. The designed P system is a cell-like P system with the nested structure of two layers and its evolution and communication mechanisms are used to realize the fuzzy clustering algorithm. This cell-like P system can effectively find the optimal cluster centers. The proposed fuzzy clustering algorithm is evaluated on a real-life gene expression data set and is further compared with the classical Iclust algorithm and the recently developed Fuzzy-VGAPS algorithm. The comparison results demonstrate that the proposed fuzzy clustering algorithm is better than other two algorithms in terms of clustering effectiveness.