Finding correlated bicluster from gene expression data of Alzheimer disease using FABIA biclustering method

Alzheimer's is a disease chronic neuro-degenerative progressive disorder in the adult human brain that tends to be old and which causes memory, thought and behavior problems and the most common causes of dementia that usually worsen over time. The microarray data of AD gene expression were taken from six different brain regions size 54675 id probes genes x 161 samples in 74 normal samples and 87 samples affected by Alzheimer's. The development of microarray technology used in dataset genetic expression includes Alzheimer's. To discover hidden pattern from microarray of Alzheimer's gene expression data we propose finding correlated bicluster using FABIA biclustering. In this paper, we apply the analysis for bicluster acquisition (FABIA) factor technique for gene expression matrix data that has been standardized median centering and normalized, and then this multiplicative method produces two very sparse Laplacian variables with heavy-tailed non-Gaussian signals for the desired bicluster. To measure information on the bicluster we use I / NI. Evaluate the results of our cycle; we use the Jaccard index and the Munkres algorithm which implemented in the Truecluster in the open source R package. From the experimental results, there are nine biclusters formed by data's AD that can be analyzed.