There is a growing interest in the use of "genetic scores" to investigate the contribution of traits such as obesity and blood pressure to disease and disease risk. However, scores developed using individuals of one ancestry (e.g. Europeans) typically are less useful when applied to populations from another (e.g. East Asians). We are investigating the performance of scores generated using data from multiple ancestries, compared with those from a single ancestry, and have found that there can be substantial improvements. As part of this, to make sure that our results are reproducible, we need to do analyses in sets of individuals who were not used in generating the scores. This project tests how our "improved" scores perform in ALSPAC mothers and fathers.