Abstract Background In order to explore the pathophysiology underlying type 2 diabetes we examined the impact of gene variants associated with type 2 diabetes on circulating levels of glucagon during an oral glucose tolerance test (OGTT). Furthermore, we performed a genome-wide association study (GWAS) aiming to identify novel genomic loci affecting plasma glucagon levels. Methods Plasma levels of glucagon were examined in samples obtained at three time points during an OGTT; 0, 30 and 120 min, in two separate cohorts with a total of up to 1899 individuals. Cross-sectional analyses were performed separately in the two cohorts and the results were combined in a meta-analysis. Results A known type 2 diabetes variant in EYA2 was significantly associated with higher plasma glucagon level at 30 min during the OGTT (Beta 0.145, SE 0.038, P = 1.2 × 10^–4) corresponding to a 7.4% increase in plasma glucagon level per effect allele. In the GWAS, we identified a marker in the MARCH1 locus, which was genome-wide significantly associated with reduced suppression of glucagon during the first 30 min of the OGTT (Beta − 0.210, SE 0.037, P = 1.9 × 10^–8), equivalent to 8.2% less suppression per effect allele. Nine additional independent markers, not previously associated with type 2 diabetes, showed suggestive associations with reduced glucagon suppression during the first 30 min of the OGTT (P < 1.0 × 10^–5). Conclusions A type 2 diabetes risk variant in the EYA2 locus was associated with higher plasma glucagon levels at 30 min. Ten additional variants were suggestively associated with reduced glucagon suppression without conferring increased type 2 diabetes risk.