Figure 1
From: Inferring neutral biodiversity parameters using environmental DNA data sets
Neutral parameter inference without dispersal limitation.
Left panels: mean MOTU rank- abundance distributions over 100 realizations for θ = 20 in a 104-read sample, without (dashed blue line) and with (black line) simulated noise: (a) 30% artifactual MOTUs added (as measured in benchmark dataset), (c) multiplicative lognormal noise of log standard deviation σlog = 1.2 (as measured in benchmark dataset), (e) multiplicative zero-truncated Poisson noise simulating barcode copy number variability (Poisson parameter λ = 4; cf. Kembel et al.52), and (g) size structure among individuals, for a ratio 
(mean body mass over birth mass). Right panels: mean and standard deviation over 100 realizations of the relative bias on the θ estimate in a 104-read sample, for θ = 1 (green), θ = 20 (black) and θ = 500 (red), as a function of (b) the proportion of artifactual MOTUs (dashed blue line underlines the linear dependence), (d) the lognormal noise intensity σlog, (f) the Poisson parameter λ, and (h) the ratio 
.
