Applying Fuzzy Iterativeself-Organizing Algorithm Based On PCA To Indentify The Candidate Genes Of Alzheimer’s Disease

Alzheimer’s disease (AD) is a dangerous dementia which may b ring great harmfulness to patients. Up to now, its cause and pathogenesis have been not clear. To the researches that try to find the way of curing AD effecti vely, it’s significantly to identify the candidate genes which are related to AD and study these genes’ pathogenesis mechanism. Because of the massive volum e and high dimensional data, we chose the computing methods to mine this da ta.Clustering analysis is a common method for processing biological data. Combining the theory of biological information, we chose improved fuzzy C-means clustering algorithm (FCM)—Fuzzy Iterative Self-Organizing Algorithm based on PCA to process the gene expression data. Firstly, the principal component analysis (PCA) algorithm is used to get the initial clustering numbers and centers of FCM. Secondly, fuzzy objective function and fuzzy constraint function are used to determine the fuzzy index of FCM. Finally, the iterative self-organizing data analysis methods (ISODATA) algorithm is introduced to optimize and control the number of categories of the improved FCM algorithm for decreasing the distortion and obtaining an accurate clustering result.By the improved clustering algorithm and the given criterion, we successfully recognized 29 genes which are virulence genes possibly. Analyzing these genes we get the following conclusions:①the abnormal expression of auxiliary protein may induce AD by inducing the abnormality of ATP;②The abnormal expression of gene RERE and HNRNPC may induce the apoptosis of nerve cells, which relate to AD.③ The decreased expression level of selenium proteins may induce AD by affecting the function of selenium.④The increased expression level of PRRG1 and Agtrl can lead to calcium dyshomeostasis, which may induce AD. The article gave the possible pathogenesis of the candidate genes of AD finally.