Fig. 5

Relative abundance assessment. a A linear model of the relationship between the expected and observed relative abundance. The dashed gray line indicates the expected 1-to-1 relationship. The plot is split by individual and bioinformatic pipeline indicated by line color. A negative binomial model was used to calculate an average relative abundance estimate across PCR replicates. To highlight quantitative performance for higher abundance features, points with observed and expected relative abundance values less than 1/median(total abundance) were excluded from the plot. b Relative abundance error rate (|expected - observed|/expected) distribution by individual and pipeline