Polygenic association in ADHD for the prediction of cognitive performance

Student thesis: Master's ThesisMaster of Science


Attention-deficit hyperactivity disorder (ADHD) is associated with multiple cognitive performance deficits. Among these Kuntsi et al (2010, Archives of General Psychiatry) selected the most promising indicators for a multivariate familial factor analysis. Two familial cognitive impairment factors were identified, which accounted for 85% and 13% of the familial variance on ADHD. The first, large factor captured speed and variability of reaction times (RT) and the second factor captured commission (Ce) and omission (Oe) errors on a go/no-go task. Using the same sample we now evaluate whether polygenic signals for ADHD generated from genome-wide SNP data are associated with the most promising cognitive variables. Analysis used data from the International Multi-Centre ADHD Gene (IMAGE) project. Transmission disequilibrium test (TDT) was conducted and the data used to generate polygenic scores in a training sample at a range of different inclusion thresholds. The best performing threshold (10% of SNP associations) was selected using cross-validation and used produce a polygenic signal which was then apply to predict cognitive performance using regression analysis in an independent test set. This study was not able to detect a positive polygenic signal for ADHD, or able to predict cognitive performance using this polygenic signal. Power calculations indicated that the training sample was insufficient in this pilot study, but demonstrated that the generation of a positive polygenic signal for ADHD should be achievable using currently available ADHD GWA samples. The study also demonstrates the effectiveness of case/pseudo-control imputation for dealing with major allele over-transmission bias, as well as suggesting a number of methodological improvements and directions for future polygenic investigations.
Date of Award2011
Original languageEnglish
Awarding Institution
  • King's College London
SupervisorJonna Kuntsi (Supervisor) & Grainne McLoughlin (Supervisor)

Cite this