As part of our analyses of prenatal nutrition we plan to analyse the effects of prenatal alcohol exposure on childhood atopic outcomes. We propose, in addition to analysing the maternal questionnaire data on alcohol intake, to utilise a Mendelian randomisation approach in order to remove potential bias and confounding and add rigour. We would like to analyse the maternal ADH1B variant (data requested) which has been shown to predict alcohol intake in pregnancy in ALSPAC. A DTA covering analysis of genetic data is already in place with QMUL. We have all other variables (maternal exposure, confounders and childhood atopic outcomes) needed to carry out these analyses.