Fig. 1.

Habitat-specific taxonomic weights classify sequences at the species level with the same confidence that uniform weights classify at the genus level (single-sided paired t-test, species bespoke vs genus uniform, t = 1.6, P = 0.14). Overlaid columns show average proportion of incorrectly classified reads for three taxonomic weighting strategies at each taxonomic level. Bespoke weights were habitat-specific. Average weights were averaged across 14 EMPO 3 habitats. Uniform weights are the current best practice. Box plots are across 14 EMPO 3 habitats. Box bounds and centre lines show first and third quartiles and median. Whiskers extend to measurements no further than 1.5 times the interquartile range from the nearest quartiles. Outliers are plotted individually. 21,513 empirical taxonomic abundances contributed to the results. Source data are provided as a Source Data file