Computational Systems Biology Methods and Protocols.7z

minimize biological and technical bias impact on research results,

give correct answers to the questions raised, and generate reason-

able hypothesis. The experiment design is determined by the

research goal (e.g., identification of diagnosis biomarker, influence

of drug treatment, outcomes of nutrition intervention), which

includes the following four important considerations: (1) experi-

ment strategy (untargeted or targeted strategies), (2) experimental

subjects (e.g., model animals or human samples, plant, cells, or

microbes), (3) biological replicates, and (4) data analytical strategy

(e.g., case-control analysis, paired analysis, time series analysis).

Untargeted and targeted strategies are two distinct experiment

strategies, which are applied in metabolomics studies widely. Each

strategy has its own inherent advantages and limitations. The untar-

geted metabolomics aims to detect metabolites as many as possible

in a biological sample. Thus in untargeted strategy metabolomics,

samples should be less preprocessed to keep more metabolites. This

strategy can identify patterns or fingerprints of metabolites

responding to genetic alterations and environmental stimuli,

which often used for hypothesis-generating and hypothesis-

refining study [6]. However, there are some difficulties of this

strategy in comparing investigation results of different laboratories,

combining research data in several time points, and characterizing

Fig. 1Flow chart of a metabolomics study

Metabolomics: A High-Throughput Platform for Metabolite Profile Exploration 267