ABSTRACT
The definition of the research question will also direct the choice of the most suitable experimental design and, consequently, will shape the data analysis plan, which is a critical part of metabolomics. A common type of research question consists of comparing groups of samples, where a group can be defined as a set of biological samples with an important characteristic in common. The last point is particularly critical, and it is also one of the weaknesses of many large-scale metabolomics research projects for biomarker discovery. Estimating, a priori, how many samples should be analyzed to arrive at a particular level of certainty is often difficult if impossible. Experimental Design is the field in statistics aiming to maximize the efficiency of experiments: as much information as possible is obtained with the least effort. The peak list will be then the output of a specific software and to exactly reproduce it, it would be mandatory to use the very same software.
