Self-organising maps (or neural networks) are analytical approaches that seek to describe structure in otherwise complex data sets. Proton NMR based lipidomics and metabolomics are just such phenotypes and are measured at multiple time points and across generations in ALSPAC (a unique resource). The proposed work will aim to look at the impact of body mass index (BMI) on the metabolome (though both observational and genetic analyses), but to characterise this impact using self-organising maps. This is an approach currently being developed by our close collaborators (Johannes Kettunen and Mika Ala-Korpela at the University of Oulu) and one we wish to employ in light of the properties of the ALSPAC data and in light of future work concerned with the same basic problem but with the added complication of metabolic challenge.