Fig. 1
From: SweepCluster: A SNP clustering tool for detecting gene-specific sweeps in prokaryotes
Outline of the clustering procedures of SweepCluster. A non-synonymous or upstream regulatory SNP is randomly chosen in each gene/operon and serves as an anchoring SNP for an initial cluster. The initial cluster is then extended by scanning and merging the neighboring SNPs or clusters. If the total span is shorter than the specified sweep length, then the surrounding SNPs or clusters are merged. Otherwise, the initial cluster is extended by merging the neighboring SNPs or clusters such that the normalized root-mean-square of inter-SNP distances (NRMSD) is minimized. All clusters after merging are re-examined and split if any inter-SNP distance within the cluster is longer than a given inter-SNP distance threshold (max_dist)