Fig. 5From: Multi-factorial examination of amplicon sequencing workflows from sample preparation to bioinformatic analysisDetection of low abundance contaminants varied among bioinformatic processing programs. Contaminants were found primarily in mothur processed data and could largely be traced back to samples corresponding to the V3-V4 primer set or to negative controls. Data shown represent total read abundance for all inexact sequence variants collapsed into a consensus sequence within 15% distance. Consensus sequences were considered mismatches were if they clustered with a 16S rRNA sequence from the mock community. White heatmap cells indicate the consensus sequence was not represented among the corresponding data. The dendrogram represents a UPGMA tree constructed from the distances among consensus sequencesBack to article page