Extended Data Fig. 1: Genome-wide association meta-analysis results for AITD in Iceland and the UK Biobank and network analysis of candidate genes.

a, Manhattan plot illustrating the findings. In total, 30,234 individuals with AITD and 725,172 controls were included in a logistic regression analysis, assuming a multiplicative model adjusted for year of birth, sex and origin (Iceland) or the first 40 principal components (the UK Biobank). Sequence variants (n = 42,907,111) were split into five classes based on their genome annotation, and the significance threshold for each class was adjusted for the number of variants in that class (Methods). In total, 93 loci passed the significant thresholds (Extended Data Table 1), whereas the red line on the plot represents a P value of 10⁻⁸. Coding variants are marked in orange and loss of function variants in red. b, Network analysis based on 37 candidate genes (outlined in Fig. 1, here marked in red), where AITD lead or correlated variants (r² > 0.8) affect protein-coding of genes (Extended Data Table 1) or mRNA expression (top cis-eQTL) in whole blood or adipose tissue (Extended Data Table 2). The network illustrates that the proteins coded by 24 of the 37 candidate genes have experimental evidence for direct interactions (blue lines) or indirect interactions (grey dotted lines, for example, one affecting the level of another), which supports a biological connection. The network analysis was performed with the IPA software (QIAGEN, https://www.qiagenbioinformatics.com/products/ingenuitypathway-analysis).