In this project I will examine the association between Skin tone, Vitamin D levels and a range of health outcomes. It has been shown in previous studies using ALSPAC data (Bonilla et al 2014) that children who had genetic variants associated with fairer skin have higher vitamin D levels. I wish to test the hypothesis that children with fairer skin will have lower BMI and blood pressure once the vitamin D levels have been controlled for. In this study I will use Instrumental Variables analysis to examine the effect of Skin tone and Vitamin D levels on a range of health outcomes such as BMI and blood pressure using the genetic variants associated with skin tone and Vitamin D levels to control for potential confounding in this estimation. Due to potential confounding in this study I will also control for time spent in the sun and whether or not a child has ever been sunburnt using questionnaire results, the gender of the child and the ethnicity of the child.

In this project I would also like to use the genotype data for the mothers to test the association between the fairer skin/higher vitamin D levels and the outcomes BMI and Blood pressure in female adults using Instrumental Variable Analysis and two sample Mendelian Randomisation. This analysis will test the same hypothesis as the analysis of the child data however rather than using Instrumental Variable analysis to assess the effect of skin tone on the outcomes considered I will use two sample Mendelian Randomisation to test the hypothesis that the SNPs shown to be associated in other studies with skin tone have an effect on BMI and Blood pressure. To control for additional confounding in this analysis I will also control for the age and ethnicity of the mother.

Levels of Vitamin D in individuals vary over the year and so in each of the analyses above I intend to control for the time of year at which the Vitamin D levels where sampled in order to control for any variation caused by differences in the season of sampling. Another source of potential confounding in this data is from population stratification due to systematic differences in the frequency of alleles for skin tone across different sections of the population. Therefore, I will control for additional potential population stratification in the data using Principal Components analysis of the genotype data.

I would like to use ALSPAC for this analysis due to the availability of the phenotypes and genotypes needed for this analysis in both children and mothers enabling analysis of the relationship between skin tone, Vitamin D and BMI/blood pressure in both children and adults.