Fig. 2: Performances of iCAMP and other approaches with simulated communities.

a–c The performances of different approaches in quantifying ecological stochasticity under high-phylogenetic-signal (HPS), medium-phylogenetic-signal (MPS), and low-phylogenetic-signal (LPS) scenarios, which were assessed by quantitative accuracy (qACC) and precision (qPRC). Data are presented as mean values ± SD. Error bars indicate standard deviations based on n = 1000 times bootstrapping from 15 independent situations. d–h The performances of iCAMP (n = 4140 comparisons = 276 comparisons among 24 biologically independent samples in each of 15 situations) and QPEN (n = 60 groups = 4 groups of comparisons among 24 biologically independent samples in each of 15 situations) were also evaluated based on the consistency between the estimated and expected relative importance of different individual ecological processes under MPS scenario. i Overall performance of iCAMP (n = 4140) and QPEN (n = 60) under MPS scenario assessed with six performance indexes: qACC (χ), qPRC (ρ), qualitative accuracy (ACC), qualitative precision (PRC), sensitivity (SST), and specificity (SPC). QPEN quantifying assembly processes based on entire-community null model analysis; NP abundance-weighted neutral taxa percentage; tNST and pNST taxonomic and phylogenetic normalized stochasticity ratio. One-side significance based on bootstrapping test (n = 1000) was indicated by a–c different letters (p < 0.05, see Source Data file for exact p values), or i ***p < 0.001; **p < 0.01; *p < 0.05 (p = 0.130, 0.000, 0.023, 0.007, 0.008, 0.031 for qACC, qPRC, ACC, PRC, SST, SPC, respectively). Source data are provided as a Source Data file.