Figure 5.
Effect of method and dataset on variant discovery performance. Sensitivity, precision and F1 score are compared for each dataset (1000 GP and SGDP) according to the type of insertion (non-reference vs reference) and the genotyping method used (1000 GP, MELT2.1.4 and TypeTE). Error bars: 95% confidence interval. Non-overlapping intervals denotes a significant difference between scores.