Sample size and statistical power considerations in high-dimensionality data settings: a comparative study of classification algorithms
1 BG Medicine, Inc., 610 Lincoln St., Waltham, MA 02451, USA
2 Institute for Bioinformatics and Translational Research, UMIT, Eduard Wallnoefer Zentrum 1, 6060 Hall in Tyrol, Austria
3 Optimal Medicine Ltd., Warwick Enterprise Park, Wellesbourne, Warwick CV35 9EF, UK
4 Division of Biostatistics and Epidemiology, University of Massachusetts - Amherst, 715 North Pleasant Street, Amherst, MA 01003, USA
BMC Bioinformatics 2010, 11:447 doi:10.1186/1471-2105-11-447Published: 3 September 2010
Data generated using 'omics' technologies are characterized by high dimensionality, where the number of features measured per subject vastly exceeds the number of subjects in the study. In this paper, we consider issues relevant in the design of biomedical studies in which the goal is the discovery of a subset of features and an associated algorithm that can predict a binary outcome, such as disease status. We compare the performance of four commonly used classifiers (K-Nearest Neighbors, Prediction Analysis for Microarrays, Random Forests and Support Vector Machines) in high-dimensionality data settings. We evaluate the effects of varying levels of signal-to-noise ratio in the dataset, imbalance in class distribution and choice of metric for quantifying performance of the classifier. To guide study design, we present a summary of the key characteristics of 'omics' data profiled in several human or animal model experiments utilizing high-content mass spectrometry and multiplexed immunoassay based techniques.
The analysis of data from seven 'omics' studies revealed that the average magnitude of effect size observed in human studies was markedly lower when compared to that in animal studies. The data measured in human studies were characterized by higher biological variation and the presence of outliers. The results from simulation studies indicated that the classifier Prediction Analysis for Microarrays (PAM) had the highest power when the class conditional feature distributions were Gaussian and outcome distributions were balanced. Random Forests was optimal when feature distributions were skewed and when class distributions were unbalanced. We provide a free open-source R statistical software library (MVpower) that implements the simulation strategy proposed in this paper.
No single classifier had optimal performance under all settings. Simulation studies provide useful guidance for the design of biomedical studies involving high-dimensionality data.