Tests of association based on genomic windows can lead to spurious associations when using genotype panels with heterogeneous SNP densities

Table 1 Average rescaled partial areas under a receiver operating characteristic curve up to a false positive rate of 5% (pAUC05) for 1 MbT, 1 MbTw and 100 T

	1 MbT	1 MbTw	100 T	1 MbFre
\(n_{QTL} = 30\)	\(4.38^a\)	\(5.27^a\)	\(5.29^a\)	\(2.37^b\)
\(n_{QTL} = 90\)	\(1.49^b\)	\(2.68^a\)	\(2.19^a\)	\(1.03^b\)
\(n_{QTL} = 300\)	\(0.52^b\)	\(1.58^a\)	\(1.48^a\)	\(1.30^a\)

Means were obtained based on 30 replicates for each number of simulated QTL. 1 MbT = GWAS analysis with a window size of 1 Mb and a constant T for all windows; 1 MbTw = GWAS analysis with a window size of 1 Mb and window specific T; 100 T = GWAS analysis with a window size of 100 SNPs and a constant T; 1 MbFre = GWAS analysis with a window size of 1 MB using EMMAX. Different superscripts in the same row indicate a significant difference (\(p < 0.05\))

ISSN: 1297-9686