5.1 Raw Data
Preprocessing and
Metabolite
Identification of MS
Platform
For GC-MS-based metabolomics, the main procedure of data pre-
processing includes baseline filtering, peak finding, retention time
correction, peak alignment, and co-eluted peaks deconvoluting.
Currently, some popular tools are developed for GC-MS data
processing, such as MetAlign [76], metaMS [77], TargetSearch
[78], and web-based MetabolomeExpress [79] and TagFinder
[80]. Specially, TagFinder is a free software package for mass spec-
tral tags retrieving based on the tag intensity correlation within a
group of time serial samples, which can be applied for retention
index (RI) and calculation, mass spectra tag (MST) extraction, and
tag structural identification. Thus, TagFinder supports both
non-targeted fingerprinting analyses and targeted metabolite
profiling.
For LC-MS-based metabolomics, the main procedure of data
preprocessing is similar to that of GC-MS-based studies. Currently,
a range of commercial software packages for LC-MS data prepro-
cessing are developed by instrument vendors (e.g., Waters Marker-
Lynx and Progenesis QI, Agilent MassHunter, Sciex MarkerView,
and LECO ChromaTOF), while some other packages are imple-
mented by academic community (e.g., XCMS [81], MZmine
[82, 83], MetAlign [76], TagFinder [84]). Specially, the XCMS
software is a widely used R package of data preprocessing with
command line mode. To facilitate application of the XCMS pack-
age, a graphic version has been developed, which allow users to
upload and process LC-MS data online. XCMS online provides a
whole workflow for untargeted metabolomics study, which
includes feature detection, retention time correction, peak align-
ment, peak annotation, statistical analysis, and data visualization
[85]. In addition, a free R package, IPO, is further designed as
assistant tool of XCMS software, which can automatically select
parameters for data processing of the LC-HRMS platform
[86]. Specially, spectra of data-independent acquisition coupled
with tandem mass spectrometry (DIA-MS/MS) platform contain
total fragment ions of all precursors simultaneously. Therefore,
DIA-MS/MS spectra are needed to be deconvoluted for fragment
ion purification. Currently, tools of DIA-MS/MS data analysis
include MS-DIAL, MetaboDIA, and MetDIA [87–89], and they
can be utilized to quantify and identify metabolites through match-
ing the MS peaks to spectra reference library, such as Metlin,
MassBank, HMDB, NIST, ReSpect, and mzCloud, LIPID
MAPS, and LipidBlast [90].5.2 Raw Data
Preprocessing and
Metabolite
Identification of NMR
Platform
The preprocessing steps in NMR platform are described as follows:
baseline correction, spectra alignment, binning, normalization, and
scaling. Baseline correction is the first step of data preprocessing,
since baseline distortion will affect the subsequent spectral analysis.
Peak shifts between different spectra are a common phenomenon
caused by instrumental factors as well as changes of the pH, saltMetabolomics: A High-Throughput Platform for Metabolite Profile Exploration 279