Fig. 4

An example of the cluster search process. The values of the two scoring metrics \((abund_i, frac_i)\) are shown beside each node v. The search results contain clusters 4 and 5 with estimated abundance 23.8 and 9, respectively. Nodes 6 and 3 did not pass the binomial test, thereby failing to traverse their descendants. The nodes in \(Path_4\) and \(Path_5\) are colored by blue and orange, respectively. \(K_4\) shares the same orange k-mers with \(L_5\). Thus, we need to adjust \(C_4\) based on cluster 5’s estimated abundance \(A_5\) to calculate the accurate scoring metrics