Applying Harmonic Functions To Identify Candidate Genes Associated With Alzheimer’s Disease

Alzheimer disease is one of the most common forms of senile dementia. Sofar, its incidence rate has been increasing but survival rate has been decreasing.However, Alzheimer disease’s complicated pathogenesis is not known to us.Simultaneously, the effective therapeutic measures is not found. Fortunately, researchesprove that AD is related to genes, now, it’s confirmed that six genes including APOE、PSEN1、PSEN2、APP are correlative with AD. In order to treat AD fundamentally, it’snecessary to identify more genes that is related with AD. Since the great amount of geneexpression data, it is necessary to apply gene chip to identify AD genes and usecomputing methods to deal with gene expression data. However, the current computingmethods can only deal with data simply and identify AD genes just by directlycomparing the difference between data. Unexpectedly, these methods cannot identifythe real causative genes fundamentally. In order to overcome this difficulty, based on theidea that Alzheimer disease may be related with the excessive imbalances of geneexpression level.we design the way that set up harmonic function to calculate thecorrelation expression values between gene and its surrounding genes, which can avoidcomparing gene expression level directly. Analyzing the correlation value of geneexpression data and comparing to the changes of the correlation value of geneexpression data in four different stages. In this paper those genes whose correlationchanges are dramatically are AD genes. At last,61genes whose changes of correlationexpression levels are dramatically are identified. Since these genes’ correlation betweenexpressions levels are obviously, one part of these genes are related to AD directly orindirectly. Unexpectedly, two genes including gene APOE that was confirmed to be ADgenes and gene cnr1that was confirmed to be AD genes by other scientists, whichshows our algorithm is effective.