Efficient Hexagonal Gridding Method for DNA Microarray Image

Abstract

In genetics, a deoxyribonucleic acid (DNA) microarray is a useful instrument that is frequently used to track thousands of genes' expression levels simultaneously. For DNA microarray, gene expression is done through microarray spot gridding, segmentation and intensity extraction. The gridding processes identify each microarray spot location and supply their coordinates for the spot. Many microarray technologies arrange their microarray spots in in rectangular fashion. However, a hexagonal grid arrangement is also being used to increase the density of the spot per unit area. While gridding microarray arranged in rectangular fashion is straightforward, gridding microarray arranged in hexagonal fashion posed its own challenge due to the irregular structure of the spot location. To solve this problem, this paper discusses two proposed methods to perform gridding for microarrays arranged in a hexagonal fashion, namely alternate-column and diamond-shape methods. These methods have been evaluated in nine different benchmark images that represent best, typical and worst-case images. Based on our experiments, the proposed diamond-shape and alternate-column methods achieved an average accuracy of 93.6% and 92.1%, respectively. In terms of computational load, the diamond shape consumes three times less computational loads as compared to the alternate-column.