Microarray Analysis Using Multiple Feature Data Clustering Algorithms

Abstract

By using Microarray Technology, in a single experiment, one can study the function of thousands of genes in parallel. Microarrays are used in various applications like disease diagnosis, drug discovery, and biomedical research. A Microarray image contains thousands of spots and each of the spot contains multiple copies of single DNA sequence. The analysis of microarray image is done in three stages: gridding, segmentation, and information extraction. The microarray image analysis takes the spot intensity data as input and produces the spot metrics as output which are used in classification and identification of differently expressed genes. The intensity of each spot indicates the expression level of the particular gene. This paper presents multiple feature clustering algorithms which extend the single feature (pixel intensity) clustering algorithms for segmentation of microarray image. The qualitative and quantitative results show that multiple feature clustering algorithms are more efficient than single feature clustering algorithms in segmenting the spot area, thus producing more accurate expression ratio.