A Genetic Approach to cDNA Microarray Image Analysis

Microarray image analysis is a significant tool for cDNA microarrays and it is divided in two main stages: Gridding and Spot-Segmentation. Most of the available microarray image analysis tools require human intervention to specify certain landmarks on the grid, or even to precisely locate individual spots. This paper focuses on the development of an original, fully automated gridding and spot-segmentation approach based on a genetic algorithm. This approach involves three main steps: a) Preprocessing of input images by wavelet-based noise reduction and Box-Cox transformation adjustment b) Gridding the preprocessed images by detecting the rectangular regions where individual spots are placed, c) Spot-segmenting together with model-based quantificating of individual spots using a genetic algorithm. The proposed genetic algorithm searches within a multidimentional-parameter space to determine, in parallel, the parameters of multiple diffusion models that optimally fit the characteristics of possible spots. Experiments with 16-bit microarray images show that the proposed method is effective and results in higher percentage of spot detection than that of existing method.