Comparison of the Mahalanobis Distance and Pearson’s χ2 Statistic as Measures of Similarity of Isotope Patterns
- 250 Downloads
To extract a genuine peptide signal from a mass spectrum, an observed series of peaks at a particular mass can be compared with the isotope distribution expected for a peptide of that mass. To decide whether the observed series of peaks is similar to the isotope distribution, a similarity measure is needed. In this short communication, we investigate whether the Mahalanobis distance could be an alternative measure for the commonly employed Pearson’s χ2 statistic. We evaluate the performance of the two measures by using a controlled MALDI-TOF experiment. The results indicate that Pearson’s χ2 statistic has better discriminatory performance than the Mahalanobis distance and is a more robust measure.
Key wordsSimilarity statistics Isotope distributions Mass spectral data interpretation Bioinformatics Mahalanobis distance
D.V. acknowledges the support of the SBO grant ‘InSPECtor’ (120025) of the Flemish agency for Innovation by Science and Technology (IWT).
- 4.Nicolardi, S., Palmblad, M., Dalebout, H., Bladergroen, M., Tollenaar, R.A., Deelder, A.M., van der Burgt, Y.E.: Quality control based on isotopic distributions for high-throughput MALDI-TOF and MALDI-FTICR serum peptide profiling. J. Am. Soc. Mass Spectrom. 21(9), 1515–1525 (2010)CrossRefGoogle Scholar
- 13.Mahalanobis, P.C.: On the generalized distance in statistics. Proc. Natl. Inst. Sci. India 2(1), 49–55 (1936)Google Scholar
- 15.Schulz-Trieglaff, O., Machtejevas, E., Reinert, K., Schlüter, H., Thiemann, J., Unger, K.: Statistical quality assessment and outlier detection for liquid chromatography-mass spectrometry experiments. BioDataMining 2(1) Article 4 (2009)Google Scholar