Fig. 3
From: SweepCluster: A SNP clustering tool for detecting gene-specific sweeps in prokaryotes
The accuracy, sensitivity and specificity of the clustering algorithm in SweepCluster in comparison with DBSCAN. The accuracy, sensitivity, and specificity were calculated for a series of values of sweep lengths or maximum inter-SNP distances using SweepCluster. Only accuracy and sensitivity were calculated using DBSCAN due to the fact that DBSCAN classified all SNPs into clusters leaving no out-of-cluster SNPs. The accuracy and sensitivity were calculated for a series of values of eps (the maximum distance between two samples) and min_samples (the minimum number of samples in a neighbhorhood) using DBSCAN. Other parameters for DBSCAN were set as default, including metric (default = ”euclidean”), algorithm (default = ”auto”), leaf_size (default = 30), and n_jobs (default = -1)