Email updates

Keep up to date with the latest news and content from BMC Bioinformatics and BioMed Central.

Open Access Methodology article

Nonparametric relevance-shifted multiple testing procedures for the analysis of high-dimensional multivariate data with small sample sizes

Cornelia Frömke1*, Ludwig A Hothorn2 and Siegfried Kropf3*

Author Affiliations

1 Department of Biometry, Hannover Medical School, Carl-Neuberg-Str. 1, D-30625 Hannover, Germany

2 Institute of Biostatistics, Leibniz University of Hannover, Herrenhäuserstr. 2 D-30419 Hannover Germany

3 Institute for Biometry and Medical Informatics, Otto von Guericke University Magdeburg, Leipziger Str. 44, D-39120 Magdeburg, Germany

For all author emails, please log on.

BMC Bioinformatics 2008, 9:54  doi:10.1186/1471-2105-9-54

Published: 27 January 2008

Abstract

Background

In many research areas it is necessary to find differences between treatment groups with several variables. For example, studies of microarray data seek to find a significant difference in location parameters from zero or one for ratios thereof for each variable. However, in some studies a significant deviation of the difference in locations from zero (or 1 in terms of the ratio) is biologically meaningless. A relevant difference or ratio is sought in such cases.

Results

This article addresses the use of relevance-shifted tests on ratios for a multivariate parallel two-sample group design. Two empirical procedures are proposed which embed the relevance-shifted test on ratios. As both procedures test a hypothesis for each variable, the resulting multiple testing problem has to be considered. Hence, the procedures include a multiplicity correction. Both procedures are extensions of available procedures for point null hypotheses achieving exact control of the familywise error rate. Whereas the shift of the null hypothesis alone would give straight-forward solutions, the problems that are the reason for the empirical considerations discussed here arise by the fact that the shift is considered in both directions and the whole parameter space in between these two limits has to be accepted as null hypothesis.

Conclusion

The first algorithm to be discussed uses a permutation algorithm, and is appropriate for designs with a moderately large number of observations. However, many experiments have limited sample sizes. Then the second procedure might be more appropriate, where multiplicity is corrected according to a concept of data-driven order of hypotheses.