Fig. 9From: A statistical framework for detecting mislabeled and contaminated samples using shallow-depth sequence dataComparing results from complete-linkage hierarchical clustering and the proposed method. Above are three two-way contingency tables comparing the results from complete-linkage hierarchical cluster analysis and the proposed method for IITA (a), NaCRRI (b), and NRCRI (c). Conflicts between the two methods are shown in red. The 146 genotypes shared between IITA and NRCRI (Fig. 4; black) are represented twice in our results: once with the 329 genotypes unique to IITA and once with the 27 genotypes unique to NRCRIBack to article page