Microarray image analysis

Abstract

Microarray is an important tool and powerful technique that is used to analyse the expression of DNA in organisms for large scale gene sequences and gene expressions. Microarray technology allows massively parallel, high throughput profiling of gene expression in a single hybridization experiment. Processing of microarray images provides the input for further analysis of the extracted microarray data. It includes the following stages: and#61623; Gridding is the process of identifying the area within an image that contain a single spot, subgrid and row, column within that subgrid. Segmentation is the process of grouping the pixels in a spot with similar features (this step results in the separation of foreground and background pixels). Intensity extraction calculates red and green foreground fluorescence intensity pairs. Preprocessing A major factor that complicates the task of image analysis and data mining is that microarray experiments involve a number of error-prone steps (occurring during fabrication, target labeling and hybridization), which induce noise on the resulting images. Microarray images are also corrupted by irregularities in the shape, size, and position of the spot. Most of the methods proposed by researchers have either considered high SNR (signal-to-noise ratio) images or various assumptions on factors such as type of thresholding used, parametric assumptions and decomposition levels. Also, some of the methods have discussed only impulse, Gaussian noise and fluorescent noise. In this research work, methods are developed, for low SNR images and estimate many other types of noises other than Gaussian, impulse and fluorescent noises. Denoising of microarray image is an essential and challenging task in the pre-processing step of microarray image analysis. newlineHere, we have proposed two novel techniques for image restoration.