GOYA STUDY

The GOYA study consists of a cohort of 5373 Danish male and female 'BMI selected individuals' with genome-wide SNP data. 2633 of these individuals were selected because they are classed as obese (BMI greater than 31 for males and females with the largest residual BMI after regression on age and parity (~BMIgreater than 33)). The remaining 2740 individuals were randomly selected from the normal distribution of BMI from the same sources. Our initial strategy used this extreme BMI selected population to identify BMI-associated SNPs across the genome. We carried out both a case/control association and a quantitative BMI association using a reverse-regression method to account for the selection bias.

We selected 5558 SNPs with pless than 0.001 (excluding those already known from previous BMI GWAS) for in-silico replication in the IARC study (n=4250).

We then took forward SNPs to genotyping replication in a further Danish population and according to three strategies:

- The top SNP from a region which had pless than 0.01 in IARC (and same direction of effect)

- The top SNP from a region with pless than 5x10-5 in GOYA, the same direction of effect in IARC and with evidence of candidacy (as assessed by Genesniffer)

- The top SNP from a region which had pless than 1x10-5 in GOYA and the same direction of effect in IARC

This gave us 28 SNPs to replicate.

We have already carried out replication in a Danish cohort, but require additional replication.

REPLICATION IN ALSPAC

We wish to test these 28 SNPs in the ALSPAC 15 year olds using data from the imputed genome-wide data.

We wish to calculate sex-specific BMI z-scores and carry out linear regression of z-score BMI on each SNP (adjusting for age).