Abstract-Spotted cDNA microarray data analysis suffers from various problems such as noise from a variety of sources, missing data, inconsistency, and, of course, the presence of outliers. This paper introduces a new method that dramatically reduces the noise when processing the original image data. The proposed approach recreates the microarray slide image, as it would have been with all the genes removed. By subtracting this background recreation from the original, the gene ratios can be calculated with more precision and less influence from outliers and other artifacts that would normally make the analysis of this data more difficult. The new technique is also beneficial, as it does not rely on the accurate fitting of a region to each gene, with its only requirement being an approximate coordinate. In experiments conducted, the new method was tested against one of the mainstream methods of processing spotted microarray images. Our method is shown to produce much less variation in gene measurements. This evidence is supported by clustering results that show a marked improvement in accuracy.

From its conception, bioinformatics has been a multidisciplinary field which blends domain expert knowledge with new and existing processing techniques, all of which are focused on a common goal. Typically, these techniques have focused on the direct analysis of raw microarray image data. Unfortunately, this fails to utilise the image's full potential and in practice, this results in the lab technician having to guide the analysis algorithms. This paper presents a dynamic framework that aims to automate the process of microarray image analysis using a variety of techniques. An overview of the entire framework process is presented, the robustness of which is challenged throughout with a selection of real examples containing varying degrees of noise. The results show the potential of the proposed framework in its ability to determine slide layout accurately and perform analysis without prior structural knowledge. The algorithm achieves approximately, a 1 to 3 dB improved peak signal-to-noise ratio compared to conventional processing techniques like those implemented in GenePix w when used by a trained operator. As far as the authors are aware, this is the first time such a comprehensive framework concept has been directly applied to the area of microarray image analysis.

No abstract

In this work, a scalable human-in-the-loop decision support system has been built around an active learning algorithm operating on time series data. Anchored in big data analytics, the system integrates an architecture component, hierarchical clustering, a random data access module, active learning, a communication and user interaction modules.

No abstract