Title: Analysing extremely small sized ratio datasets

Authors: Piero Ricchiuto; Judy C.G. Sng; Wilson Wen Bin Goh

Addresses: Cardiovascular Division, Harvard Medical School, 25 Shattuck St Boston, MA 02115, USA ' Singapore Institute for Clinical Sciences, Brenner Centre for Molecular Medicine, 30 Medical Drive, 117609 Singapore; Department of Pharmacology, National University of Singapore, MD11 CRC 05-34, 117597 Singapore ' Cardiovascular Division, Harvard Medical School, 25 Shattuck, St Boston, MA 02115, USA; School of Computing, National University of Singapore, COM1 13 Computing Drive, 117417 Singapore

Abstract: The naïve use of expression ratios in high-throughput biological studies can greatly limit analytical outcome especially when sample size is small. In the worst-case scenario, with only one reference and one test state each (often due to the severe lack of study material); such limitations make it difficult to perform statistically meaningful analysis. Workarounds include the single sample Z-test or through network inference. Here, we describe a complementary plot-based approach for analysing such extremely small sized ratio (ESSR) data - a generalisation of the Bland-Altman plot, which we shall refer to as the Dodeca-Panels. Included in this paper is an R implementation of the Dodeca-Panels method.

Keywords: proteomics; bioinformatics; clinical translation; drug discovery; biomarker discovery; small ratio datasets; protein responders.

DOI: 10.1504/IJBRA.2015.069225

International Journal of Bioinformatics Research and Applications, 2015 Vol.11 No.3, pp.268 - 280

Received: 20 Aug 2014
Accepted: 07 Dec 2014
Published online: 05 May 2015 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article