Introduction
In recent years not only subclinical thyroid dysfunction, but even subtle variation in thyroid function within the normal range has been associated with various clinical complications. Twin studies have estimated that 57-71% of the variation in thyroid function is explained by genetics. Here, we report the largest genome-wide association study (GWAS) meta-analysis for normal thyroid function to date.

Methods
We conducted GWAS meta-analyses for thyroid function in 271,027 individuals without thyroid disease from 45 cohorts collaborating in the ThyroidOmics Consortium. Next to TSH, fT4 and subclinical hypo- and hyperthyroidism, we included novel traits: free T3 (fT3), total T3 (TT3), T3/fT4 ratio (i.e., a marker for thyroid hormone metabolism), and TSHxfT4 product (marker for the hypothalamic-pituitary-thyroid axis setpoint). Subsequent analyses included colocalization analyses with GTEx data and Mendelian Randomization (MR) assessing causality with thyroid cancer. 

Results
In total, 525 independent genetic variants were identified: 259 for TSH (189 novel), 85 for fT4 (63 novel), 9 for fT3, 17 for TT3, 18 for fT3/fT4, 4 for TT3/fT4, 94 for TSHxfT4, and 31 and 8 for hyper- and hypothyroidism, respectively. While several genes (e.g. TSHR, TPO, HLA-C, and SH2B3) have an established role in thyroid hormone regulation or autoimmune thyroid disease, the role of most loci is still unknown. Colocalization analyses yielded associations of mRNA levels of 237 genes with TSH in predominantly thyroid, and 146 genes with fT4 in peripheral tissues. MR showed a causal relation between thyroid function and the risk of thyroid cancer. 

Conclusion
These results do not only substantially improve our understanding of the genetic basis of thyroid function, but also pave the way for future studies integrating these markers in the clinical management of thyroid disease patients.