MetabolAnalyze

Version: 1.3

Website

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

Description

This CRAN package provides probabilistic PCA (PPCA), probabilistic principal components and covariates analysis (PPCCA) and mixtures of probabilistic principal component analysis (MPPCA) for modelling metabolomics data. PPCA is based on the Gaussian latent variable model. PPCCA incorporates covariates into the model, which is useful for analysing human data, where other phenotypic data (age, gender, BMI) are also gathered, as well as metabolomic data. In this model it is assumed that the latent variable distribution for observations follows a multivariate Gaussian distribution that is not centred at the origin, like PPCA. A collection of PPCA models can be combined for MPPCA for non-linear modelling purposes. MPCCA clusters observations into groups and reduces data dimensionality.

Functionality

• 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

mclust, mvtnorm, ellipse, gtools, gplots

Input Formats - Open

Peaklist matrix

Input Formats - Proprietary

N/A

Published

2010

Last Updated

2015

License

GPL-2

Paper

https://www.ncbi.nlm.nih.gov/pubmed/21092268

PMID

21092268