pathophysiological stimuli or genetic modification [1, 2]. From the
view of genetic information flow, small molecular metabolites
locate in the downstream of genes and proteins and correlate to
the phenotype closely. A minor change of gene expressions or
protein modifications upstream might result in the readily amplifi-
cation of metabolites downstream. And genomic study may not
directly indicate what may happen, while metabolomics research
may depict what has happened. Thus, metabolomics has the poten-
tial as a more reliable and sensitive measuring approach in early
diagnosis of disorder for organism and plays an important role in
bridging the gaps between genotype and phenotype. Metabolomics
covers a wide range of metabolites with heterogeneous structures
and broad dynamic concentration change of compounds in a
biological system [3]. Nowadays, over 72,619 small-molecule
metabolites and their concentration changes in biological systems
are described in the Human Metabolites Database (HMDB:
http://www.hmdb.ca/, version 3.6). Metabolic profile in a
biological system covers a wide range of metabolites with heteroge-
neous structures and their dynamic range of concentrations across
nine orders of magnitude in a biological system; thus a comprehen-
sive detection of metabolic profile for a biological system is a great
challenge [4]. The combination of multiple detecting platforms of
metabolomics can enhance analysis of such a broad spectrum of
metabolites. In practice, numerous detecting platforms have been
utilized, for example, nuclear magnetic resonance (NMR), gas
chromatography (GC) coupled with mass spectrometry (MS), and
ultrahigh-performance liquid chromatography (UHPLC) coupled
with mass spectrometry (MS) [5]. Compared with NMR methods,
MS-based approaches exhibit a higher sensitivity for metabolite
detection. Typically, GC and LC coupled to MS are employed
widely since these methods combine an effective chromatographic
separation with a sensitive mass detection. A classical research road-
map of metabolomics is illustrated in Fig.1 for both untargeted and
targeted strategies, which includes section of experimental design,
sample collection, sample preparation, data acquisition, data pre-
processing, data analysis, mechanism interpretation, and hypothesis
generation. In this review we will discuss these different sections of
metabolomics.2 Experimental Design
Metabolomics is a powerful “omic” tool in the study of many
research fields ranging from biomarker identification to disorder
mechanism interpretation. One of the purposes for metabolomics
study is to quantitatively measure small-molecule metabolites in
biological samples, which contain complex compounds and have
complicate bias. An appropriate experimental design should266 Jing Cheng et al.