All analyses are performed in a sample of Dutch families from the Netherlands Twin Register (NTR). To maximize statistical power, imputation of genotypes in the NTR sample was based on two alternative reference panels: the 1000G Phase 1 project reference panel (Consortium 2012) and the reference panel generated by the Genome of the Netherlands (GoNL) project (Boomsma et al. 2014; The Genome of the Netherlands 2014). The GoNL reference panel was derived by sequencing the whole genome of 250 trio-Dutch families and matches therefore the ancestral background of our sample. The GoNL panel is expected to facilitate imputation of variants which are specific to the Dutch population (Boomsma et al. 2014). Furthermore, the use of the GoNL panel is expected to result in higher imputation accuracy relative to the 1000G panel, especially for low frequency GVs (MAF < 5 %) (The Genome of the Netherlands 2014). Such increased accuracy is expected to increase the statistical power to capture the signal in the measured GVs.
