Fig. 3

Quantitative ability for mOTUs2 and Kraken. mOTUs2 output reads vs true reads (a) and Kraken output reads vs true reads (b). For all true-positive genera classifications (Spearman’s rank correlation coefficients R²=0.91 and R²=0.69, for n = 2084 and n = 2021 true-positive classifications for mOTUs2 and Kraken, respectively). All 100 simulated datasets were first quality trimmed using Trimmomatic and depleted for human reads using the best parameters as previously mentioned. mOTUs2 classifications were left unfiltered whereas Kraken had a confidence threshold of 0.2 and a minimum read threshold of 500 applied