Figure 3.

Mean predictive accuracy in the test data sets. Each data point represents the mean of 100 trials. Error bars depict 95% confidence intervals. Predictive accuracy increases with the number of SNPs used, but is still very good even with only a small number of SNPs.

Allocco et al. BMC Genomics 2007 8:68   doi:10.1186/1471-2164-8-68
Download authors' original image