muma

Version: 1.4.0

Website

https://cran.r-project.org/web/packages/muma/index.html

Description

Metabolomics Univariate and Multivariate Analysis (muma - a CRAN package) provides a pipeline for analysis, including preprocessing, statistical analysis and metabolite identification. It also specifically provides a number of analysis tools for NMR data - statistical total correlation spectroscopy (STOCSY) and ratio analysis of NMR spectroscopy (RANSY). Initially a dataset is read and then screened for missing values. Four imputation methods are provided: mean, minimum, half-minimum and zero. Normalisation is then performed for the whole spectra, every variable is transformed into a fraction of the total intensity of the spectrum. A range of scaling techniques are then available: Auto, Pareto, Range, Vast and Medium, with each variable being mean centred before scaling. It provides a variety of both univariate (Welch's T test and Wilcoxon-Mann Whitney) and multivariate (PCA, PLS-DA and OPLS-DA) statistical techniques, as well a method of evaluating the clustering power of PCA. For univariate analysis a decision tree algorithm performs hypothesis tests on a single variable - first Shapiro Wilk's test is used to assess whether the data are normally distributed or not, then either Welch's T test or Wilcoxon-Mann Whitney test are performed. Benjamini-Hochberg multiple testing correction can then be applied. Correlation heatmaps can be created to visualise the data. With NMR data, these heatmaps are referred to as STOCSY and correlations >0.95 indicate a structural relationship between peaks belonging to the same molecule. RANSY is also designed to show the structural relationships between peaks.

Functionality

• Post-processing
• Statistical Analysis

Instrument Data Type

• MS
• NMR

Approaches

Computer Skills

Advanced

Software Type

R Package

Interface

Command line interface

Operating System (OS)

• Unix/Linux
• Mac OS
• Windows

Language

R

Dependencies

car, pdist, pls, gplots, mvtnorm, robustbase, gtools, bitops, caTools, pcaPP, rrcov

Input Formats - Open

.csv

Input Formats - Proprietary

N/A

Published

2012

Last Updated

2015

License

GPL-2

Paper

http://www.eurekaselect.com/107837/article

PMID