Figure 5

Putative, missing and spurious hotspots for the methods applied to the yeast dataset generated in 2005 [2]. (a) The average over all genes of the − log of the maximum P value of the two yeast datasets for each SNP. (b)-(g) The average over all genes of the − log P value for each SNP for several methods: the standard t-test, SVA, ICE, LMM-EH, PANAMA and NICE. Blue asterisks show putative genetic regulatory hotspots predicted from the merged dataset, green arrows show missing hotspots and red arrows show spurious hotspots identified by each method. Red horizontal lines show thresholds used to select significant peaks, which are two standard deviations above the mean. Note that the t-test has a distinct advantage in this evaluation because P values from the t-test were used to determine the putative regulatory hotspots.