A novel neural network approach to cDNA microarray image segmentation

Abstract

Microarray technology has become a great source of information for biologists to understand the workings of DNA which is one of the most complex codes in nature. Microarray images typically contain several thousands of small spots, each of which represents a different gene in the experiment. One of the key steps in extracting information from a microarray image is the segmentation whose aim is to identify which pixels within an image represent which gene. This task is greatly complicated by noise within the image and a wide degree of variation in the values of the pixels belonging to a typical spot. In the past there have been many methods proposed for the segmentation of microarray image. In this paper, a new method utilizing a series of artificial neural networks, which are based on multi-layer perceptron (MLP) and Kohonen networks, is proposed. The proposed method is applied to a set of real-world cDNA images. Quantitative comparisons between the proposed method and commercial software GenePix are carried out in terms of the peak signal-to-noise ratio (PSNR). This method is shown to not only deliver results comparable and even superior to existing techniques but also have a faster run time.

A Systematic Examination of Microarray Segmentation Algorithms

“…Application of different algorithms on images from replicated spot experiments could result in suitable comparison between the segmentation algorithms in terms of consistency. In [22,32], the authors compare six DNA segmentation methods in terms of mean absolute error (MAE) [30,31] of replicated spots. Their study shows that the graph-based segmentation method has a smaller MAE compared to circle-shape based and region growing-based segmentation schemes..…”

Microarray spot partitioning by autonoumsly organising maps thorugh contour model

“…During a training session by selecting a appropriate number of neurons [35,36] and topology [37] for AOM the intensity SP in_tr (x t ) of arbitrarily chosen pixel x t of training image taken as AOM input at time t=0, 1,…….t max_tr. Let t max_tr be the iterating value for training of AOM.…”

Segmenting Microarrays with Deep Neural Networks

“…Neural networks have been applied to the microarray image segmentation problem previously [19], but the networks used were shallow (three layers) and applied only as part of a larger heuristic. In contrast, our inspiration was Ciresan et al [4], where a deep network was used to identify neuron membranes in electron microscopy images.…”