Skip to navigation Skip to main content Skip to footer

Approved research

Using Biobank polygenic scores to predict early cognitive development and educational achievement

Principal Investigator: Robert Plomin
Approved Research ID: 18079
Approval date: April 1st 2016

Lay summary

Early prediction of complex traits such as disease risk or educational propensity is an ultimate goal of personalised medicine and education. Although GWAS are increasingly successful in identifying trait-associated loci, due to polygenicity single-variant models have limited prediction value. We will use UK Biobank data to construct polygenic models to predict individual phenotypes in our UK-representative sample of 6000 adolescents assessed longitudinally from infancy (Twins-Early-Development-Study). We will also investigate trait-specific assortative mating in Biobank participants, and test how/if this is reflected in genetic architecture of traits in our target sample. Early prediction of complex traits such as disease risk or educational propensity is an ultimate goal of personalised medicine and personalised education. Educational attainment and cognitive ability are associated with a host of health and employment outcomes; early prediction of cognition-related problems and potential is of public interest because it is the first step towards developing prevention and intervention strategies. We will use genotypic and phenotypic data from the UK Biobank to build polygenic models that will then be used to predict individual phenotypes in our target sample by directly estimating genetic similarity amongst discovery and target sample. In particular, we will focus on cognitive, education, and socio-economic phenotypes, using polygenic scores for height, body mass index as well conditioned ?anchor? variables for comparison. To have maximum power to predict phenotypes, we require access to the full cohort, because one of the primarily limiting factors of prediction is the size of the discovery sample. Our analyses will, therefore, require individual-level imputed genotype and phenotype data.