Towards reliable microarray analysis and design


Microarray studies represent an extraordinary advance but are notoriously noisy, produce results hardly reproducible, and are therefore unreliable at best. We present a framework that affords a quantitative quantification of their inherent noise and (in)accuracy and, more importantly, a methodology to remove noise and improve their reliability and reproducibility. Furthermore, we present a principled methodology to design a variant of new generation microarray (nxh-chips) that are provably reliable by two new metrics of reliability. Potential impact is illustrated with a new approach to phylogenetics that produces a universal tree of life (TOL) based on more objective criteria and is feasibly scalable to entire genomes, while being implementable on current microarray biotechnology. A preliminary TOL is actually shown as a phylogenetic analyses of a selection of 22 species/ biomarkers from the standard Biomarker Identification Numbers (BOLD/BINs). Conclusions include some discussion of theoretical and experimental problems for implementation of this methodology in practice.

Publication Title

Proceedings of the 9th International Conference on Bioinformatics and Computational Biology, BICOB 2017

This document is currently not available here.