Data Processing, Metabolomic Databases and Pathway Analysis

Metabolomics relies on a variety of bioinformatics tools aiming to derive information from data. Data acquisition with chromatography-coupled mass spectrometry yields high volumes of raw data that need to be processed to achieve de-noised andmeaningful biochemical data that subsequently can be presented for statistics and pathway analysis in a coherent manner. A variety of new tools and algorithms have recently been developed that help in this process. Nevertheless, a surprisingly low number of metabolic signals can be unambiguously identified. If cutting-edgemethods areused, we can confidently interpret signals that can neither be identified by reference standards nor annotated by databasematching as 'novel metabolites'. Such methods may comprise high-resolution, high-accurate mass data in conjunction with good data alignment programs that perform peak picking and deconvolution, calculation of elemental formulas, database queries and constraining hit lists by interpreting mass spectral fragmentations and retention-based metabolomic libraries. In an effort to improve standardizations formetabolomic reports, we propose that not only metabolites must be named but also, for all reported metabolites, identifiers or structure codes (InChI keys) from public (bio)chemical databases need to be detailed. Machine-readable structures will vastly accelerate our knowledge of the magnitude and importance of the metabolome in various plant species, organs and physiological conditions. We detail how well-annotated metabolome data can then be used to visualize and interrogate biochemical pathways by matching information to the broad plant databases that currently exist.