Computational protocol: Investigating shared aetiology between type 2 diabetes and major depressive disorder in a population based cohort

Similar protocols

Protocol publication

[…] The list of T2D risk SNPs selected for MR was made based on evidence for prior association with T2D. Single nucleotide polymorphisms (SNPs) found to be associated (P ≤ 5 × 10−8) with T2D in two GWAS (comprising 38,840 cases and 114,981 controls [Morris et al., ] and 47,979 cases and 139,611 controls comprising a trans‐ancestry GWAS [Mahajan et al., ]) were used to perform MR by testing for their association with MDD and current psychological distress. The list consisted of 10 independently associated SNPs from DIAGRAM GWAS that were significant at a genome‐wide level [Morris et al., ], and seven further independent loci identified in the DIAGRAM trans‐ancestry T2D GWAS [Mahajan et al., ]. 11/17 SNPs were directly genotyped in GS:SFHS and these were the SNPs used in this study (Table ). These SNPs have been validated for their association with T2D using a two‐stage meta‐analyses replication within the original GWAS studies. All SNPs have been found to be associated in European populations and are therefore suitable proxies for T2D in the present study. Using a Bonferroni correction for multiple testing we calculated the threshold for statistical significance for the MR analyses to be (P < 0.0045 [0.05/11]). PLINK was used to calculate the number of minor alleles to create a variable for association testing [Purcell et al., ]. Depression‐associated SNPs were not tested for association with T2D as the only two robustly associated MDD SNPs were identified in a sample of Chinese women, and do not replicate in the largest MDD GWAS of European descent [CONVERGE, ].SNPs were tested for their association with MDD and GHQ‐28 scores in GS:SFHS using mixed linear models implemented in AS‐Reml‐R (www.vsni.co.uk/software/asreml) software package. Age, sex and SNP allele count were fixed effects. To control for relatedness between individuals family structure was fitted as a random effect by creating an inverse relationship matrix using pedigree kinship information. Wald's conditional F‐test was used to calculate the significance of fixed effects. If T2D SNPs are associated with MDD via a causal pathway involving diabetes then the association should only be present in diabetic individuals. A sensitivity analysis was carried out to determine these effects by testing for SNP association in diabetic and control individuals separately. As there were only 130 individuals in GS:SFHS with both diabetes and depression the sensitivity analysis was only performed for GHQ scores. If an association is observed in non‐diabetics then the SNPs may affect diabetes and depression independently (pleiotropy) and the assumptions of MR are violated. [...] GWAS summary statistics for the DIAGRAM T2D GWAS and the PGC MDD GWAS were used to perform LD score regression. This method uses the correlational nature of SNPs such that SNPs with high LD will have higher average χ2 statistics than those with low LD. To estimate genetic correlations the product of two z‐scores from GWAS of two traits can be regressed onto the LD score and the slope of the regression used to estimate genetic covariance [Bulik‐Sullivan et al., ]. The intercept was left unconstrained as the degree of sample overlap between DIAGRAM T2D and PGC MDD cohorts was unknown. […]

Pipeline specifications

Software tools PLINK, ASREML, LDSC
Application GWAS
Diseases Diabetes Mellitus, Diabetes Mellitus, Type 2, Metabolic Diseases