Power Analysis and Sample Size Determination in Metabolic Phenotyping

Abstract

Estimation of statistical power and sample size is a key aspect of experimental design.However, in metabolic phenotyping, there is currently no accepted approach for these tasks, in large part due to the unknown nature of the expected effect. In such hypothesis free science, neither the number or class of important analytes, nor the effect size are known a priori. We introduce a new approach, based on multivariate simulation, which deals effectively with the highly correlated structure and high-dimensionality of metabolic phenotyping data. First, a large data set is simulated based on the characteristics of a pilot study investigating a given biomedical issue. An effect of a given size, corresponding either to a discrete (classification) or continuous (regression) outcome is then added. Different sample sizes are modeled by randomly selecting data sets of various sizes from the simulated data.We investigate different methods for effect detection, including univariate and multivariate techniques. Our framework allows us to investigate the complex relationship between sample size, power and effect size for real multivariate data sets. elegans.

Nutrimetabolomics: An Integrative Action for Metabolomic Analyses in Human Nutritional Studies

“…Higher numbers of biological replicates thus allow for a better characterization of the studied population. However, in metabolomic studies, it is often not known which variables will change and the classical approaches to sample size can normally not be applied …”

“…The main advantages of this approach include the ability to determine sample size, even when experimental pilot data are not available, and technique specificity (NMR and MS), which can improve the power of the study. More recently, Blaise and colleagues developed a method for performing power calculations for metabolic phenotyping. The method is composed of three steps: i) modeling of the distribution of pilot data; ii) introducing an artificial effect; and iii) estimating confidence intervals for performance metrics.…”

Experimental Design and Sample Preparation in Forest Tree Metabolomics

“…For a stronger and significant statistical analysis, the number of replicates needed can be established by power analysis determined from the degree of analytical variance within the populations under study [50]. Statistical power analysis relates sample size, effect size (i.e., the difference of two group means divided by the pooled standard deviation) and significance level to the chance of detecting an effect in a dataset, and thus, should be performed before conducting the experiment as a key step in the experimental design [51,52]. Information for power analysis can be obtained through pilot studies or extrapolated from the literature [51].…”

“…Statistical power analysis relates sample size, effect size (i.e., the difference of two group means divided by the pooled standard deviation) and significance level to the chance of detecting an effect in a dataset, and thus, should be performed before conducting the experiment as a key step in the experimental design [51,52]. Information for power analysis can be obtained through pilot studies or extrapolated from the literature [51]. Sample size determination modules can be found in bioinformatic tools for metabolite data analysis, such as MetaboAnalyst 3.0, based on the Bioconductor R package Sample Size and Power Analysis (SSPA) and using data from a pilot metabolomic study [52].…”

Complex mixtures by NMR and complex NMR for mixtures: experimental and publication challenges

“…This is a known issue in multivariate analysis and some corrections to the p ‐values have therefore been introduced . Nevertheless, some of the relevant statistics that support these criticisms such as statistical power or sample size have not yet been fully ‘ported’ from the univariate to the multivariate world in which metabolomics reigns, although recent efforts have been made in this direction . The take‐home message of Ioannidis is that large‐scale studies must be generally preferred over multiple small‐scale studies.…”