Bioinformatic solutions for addressing the grape metabolome

A complete overview of all metabolites present in a particular organism forms an invaluable source of information, even if available in only semiquantitative or qualitative form – the human metabolome database is a good example. For other organisms, such resources are rarely available. In our laboratories we are concentrating on the elucidation of the metabolome of the grape berry, and I will describe the challenges that arise in attempting to obtain a comprehensive overview of all chemical compounds, distributed over many different chemical classes. This wide variety in metabolites forces the use of several different analytical platforms, each one necessitating its own analytical optimisation [1-3]. Conversely, data analysis is a complicated and elaborate process in which several sources of information need to be combined. The complex nature of the raw data necessitate extensive fine-tuning, and building in quality control steps along the process is vital. While commercial software, as provided by, e.g., instrument manufacturers, has shown dramatic improvements in capabilities, flexibility and quality, it is in general a black box and, more often than not, vendor-specific. Currently, open-source software forms a viable alternative, with powerful methods and a large user community. The only downside is the level of expertise required: these programs can only be used with adequate bioinformatics support. This contribution highlights the elements of our in-house data-processing pipeline for the elucidation of the grape metabolome, processing hundreds of large data files and yielding intensity values for thousands of variables. The results contain both known and unknown compounds, and serve as the starting point for more elaborate statistical analyses. Practical implications of such labour- and time-intensive scientific projects will be discussed as well