MDQC: a new quality assessment method for microarrays based on quality control reports

• DOI: 10.1093/bioinformatics/btm487
• Publication Date: 2007-10-13T00:33:48Z
• Authors: Gabriela V. Cohen...; Zsuzsanna Hollander; Enqing Shen; Ruben H. Zamar; Robert Balshaw; Andreas Scherer; Bruce McManus; Paul Keown; W. Robert McMaster; Raymond T. Ng
• Keywords: Mahalanobis distance; Bioconductor; Quality Score

Abstract Motivation: The process of producing microarray data involves multiple steps, some which may suffer from technical problems and seriously damage the quality data. Thus, it is essential to identify those arrays with low quality. This article addresses two questions: (1) how assess a dataset using measures provided in control (QC) reports; (2) possible sources problems. Results: We propose novel multivariate approach evaluate an array that examines ‘Mahalanobis distance’ its attributes other arrays. we call Mahalanobis Distance Quality Control (MDQC) examine different approaches this method. MDQC flags problematic based on idea outlier detection, i.e. whose jointly depart bulk Using case studies, show analysis gives substantially richer information than analyzing each parameter QC report isolation. Moreover, once produced, our assessment method computationally inexpensive results can be easily visualized interpreted. Finally, computing these distances subsets increase method's ability detect unusual helps reasons Availability: library implement will soon available Bioconductor Contact: [e-mail] Supplementary information: are at Bioinformatics online.