Microarray subgrid detection: a novel algorithm
International Journal of Computer Mathematics - Bioinformatics
A novel neural network approach to cDNA microarray image segmentation
Computer Methods and Programs in Biomedicine
| Hi-index | 0.00 |
In this paper, a completely blind microarray image gridding framework is developed. The only input to the framework is the microarray image, which can be at any resolution, and the gridding is accomplished with no prior assumptions. The framework includes an evolutionary algorithm (EA) and several novel methods for various stages of the gridding process including subgrid detection. The approach toward gridding differs significantly from most existing gridding frameworks as it does not make use of 1D projections at any stage. Also proposed is the concept of regular spaced grid fitness. Rather than simply trying to identify the number of rows and columns within the grid, the approach includes a measure of fitness for possible grids. By attempting to minimize this fitness value, there is a proven measure of consistency to gridding across multiple images. The framework is robust against high levels of image noise and a high percentage of nonexpressed/undetectable spots. The developed framework is thoroughly tested with a large number of simulated grids and several real microarray images.