Tools for processing and analyzing chemometric data
Traditionally, the practitioner of a chemometrics protocol is faced with the simultaneous analysis of tens to hundreds of observations. In the case of NMR metabolomics, these datasets are commonly 1D 1H NMR spectra of some cell metabolite extract or biofluid. It is not uncommon to find ad hoc toolchains or protocols involving multiple different software packages and manual data curation steps between data acquisition and data analysis. While it is still somewhat ad hoc in a broader sense, the MVAPACK toolbox provides a flexible platform in which both new and established data handling methods may be tested and automated, at least in the context of NMR datasets.
The MVAPACK toolbox provides NMR data processing functions, statistical scaling and normalization functions, model-building routines and validation functions to facilitate rapid data analysis in NMR chemometrics. The many functions in MVAPACK are documented in the MVAPACK manual, available for download here:
The MVAPACK toolbox is written in (mostly MATLAB-compatible) GNU Octave language, and is completely open-source under the GNU GPLv3. MVAPACK runs and has been tested on the GNU/Linux and Mac OS X operating systems, and requires NMRPipe to load NMR data. Feel free to use or modify the code to your purposes, and also feel free to submit bug reports and patches to our MVAPACK address.
If you find MVAPACK useful in your research, we ask that you cite the following work:
- (116) B. Worley, and R. Powers* (2015) "Generalized Adaptive Intelligent Binning of Multiway Data", Chemometr. Intell. Lab. Syst., 146:42-46 PMC4456038.
The coffees dataset, published to illustrate the use of MVAPACK on a real dataset for real inference, is available in two forms. The first form is a considerably smaller file and includes the NMR data and the MVAPACK script used to analyze it. The second form is larger, and contains the resulting GNU Octave session data of running the script on the data. Both are available below.
MVAPACK is open source and freely available for academic use. Industrial users may contact Robert Powers for licensing details.
Copyright (C) 2014 University of Nebraska Board of Regents.
Get MVAPACK
- MVAPACK is freely available for academic use. Just enter your name, email and institution on the next page, and we'll send you a download link!
- Click here to download MVAPACK.
MVAPACK Online
- A new MVAPACK web interface is now available for academic use! The web site guides you through the entire data processing pipeline without the need to know commands or the MVAPACK script language. Try it now!
- Click here to access MVAPACK online.
- Click here to register.
References
Publications related to MVAPACK
- (116) B. Worley, and R. Powers* (2015) "Generalized Adaptive Intelligent Binning of Multiway Data", Chemometr. Intell. Lab. Syst., 146:42-46 PMC4456038.
- (111) B. Worley and R. Powers* (2014) "MVAPACK: A Complete Data Handling Package for NMR Metabolomics", ACS Chem. Biol., 9(5):1138-1144 PMC4033658.
- (121) B. Worley and R. Powers* (2015) "A Sequential Algorithm for Multiblock Orthogonal Projections to Latent Structures", Chemometr. Intell. Lab. Syst., 149(Part B):33-39 PMC4668594.
- (107) S. Lei, L. Zavala-Flores, A. Garcia-Garcia, R. Nandakumar, N. Madayiputhiya, Y. Huang, E. D. Dodds,R. C. Stanton, R. Powers*, and R. Franco* (2014) "Alterations in Energy/Redox Metabolism Induced by Mitochondrial and Environmental Toxins: A Specific Role for Glucose-6-Phosphate-Dehydrogenase and the Pentose Phosphate Pathway in Paraquat Toxicity", ACS Chem. Biol., 9(9):2032-2048 PMC4168797.
- (96) B. Worley, and R. Powers* (2013) "Multivariate Analysis in Metabolomics", Current Metabolomics, 1(1):92-107 PMC44651877.
