The aim of this project is to use advanced machine learning and data mining techniques to extract features and patterns from the ALSPAC data in an unsupervised manner. Techniques to be used include subgroup discovery and exceptional model mining, which find subpopulations that are statistically robustly different wrt. properties of interest compared to the overall population. Epidemiological properties of interest include very dense genetic data (genome wide single nucleotide polymorphism, copy number variation and sequence data), gene expression and gene methylation data and outcomes related to body composition, obesity, physical activity, health related behaviours such as diet and smoking and cognitive function.