Astream

Version: 2.0

Website

http://www.urr.cat/AStream/AStream.html

Description

An R package for the annotation of preprocessed LC-MS data (a data.frame of features), detecting isotopes, adducts and fragments. xcmsSet objects can be directly imported into AStream. A quality control step is implemented to identify outlier peaks, which are then discarded. In this peaks below a given intensity are regarded as having 0 intensity and the counts for all samples are transformed into Z-scores, which are used to detect outliers with a high number of 0 intensities. Sample intensity quartiles for probabilities 0.1, 0.2, 0.3, 0.4 and 0.5 are also calculated and standardised as Z-scores. Samples with overly distinct intensity distribution profiles are then also discarded as outliers. The threshold levels for outlier detection can be modified by the user. Signal intensity is then normalised across samples. A pairwise correlation analysis across all features is then performed to identify features which are the same metabolic compound (due to fragmentation). HMDB metabolites are used for compound identification, using monoisotopic mass and biofluid/tissue location. Feature pairs corresponding to carbon isotopes of the same compounds are then identified. Adducts are then identified by searching for peaks within the m/z range of the expected m/z of the adduct.

Functionality

• Annotation/MS/Level 4 - Unequivocal Molecular Formula

Instrument Data Type

• MS/LC-MS

Approaches

• Metabolomics/Untargeted

Computer Skills

Advanced

Software Type

R Package

Interface

Command line interface

Operating System (OS)

• Unix/Linux
• Mac OS
• Windows

Language

R

Dependencies

R (≥ 3.0.1), multtest, Biobase, plotrix

Input Formats - Open

data.frame, xcmsSet

Input Formats - Proprietary

N/A

Published

2010

Last Updated

2014

License

GPL (≥ 2)

Paper

https://www.ncbi.nlm.nih.gov/pubmed/21414990

PMID

21414990