MetMSLine

Version: 1.2.1

Website

http://wmbedmands.github.io/MetMSLine/

Description

This package provides a workflow for untargeted LC-MS data, including normalisation, smoothing, biomarker selection, annotation and metabolite identification. As input this package requires a peak list. Zero-filling is first performed, then median fold-change normalisation can be applied if required. QC-based locally weighted scatterplot smoothing is then used to alleviate the effect of signal drift. Data are then log transformed and features which are found stability across pooled samples are maintained. Outliers are then removed using expansion of Hotelling's T2 distribution ellipse (samples beyond this ellipse but within first/second compound PCA plot are removed). Potential biomarkers are then identified using multivariate regression and visualised with heatmaps. Using MS/MS data, features can then be identified by matching to fragments MS/MS spectra. Metabolites can also be identified via a user-inputed .csv database of compounds and their exact masses.

Functionality

• Workflow

Instrument Data Type

• MS/LC-MS

Approaches

• Untargeted

Computer Skills

Advanced

Software Type

R Package

Interface

• Command line interface
• Graphical user interface

Operating System (OS)

• Unix/Linux
• Mac OS
• Windows

Language

R

Dependencies

R (≥ 3.2.0), fgui, tcltk2, zoo, gplots, ggplot2, sfsmisc, RColorBrewer, readMzXmlData

Input Formats - Open

Peaklist (xcms), covariate table (.csv)

Input Formats - Proprietary

N/A

Published

2014

Last Updated

2017

License

GPL-3

Paper

http://www.ncbi.nlm.nih.gov/pubmed/25348215

PMID

25348215