Fig. 8
From: A framework for assessing 16S rRNA marker-gene survey data analysis methods using mixtures.
Feature-level differential abundance assessment. Log-fold change error bias (a) and variance (b) metric distribution by subject and pipeline. The bias (1−slope) and variance (R2) metrics are derived from the linear model fit to the estimated and expected log fold-change values for individual features. Boxplot outliers, 1.5×IQR from the median were excluded from the figure to prevent extreme metric values from obscuring metric value visual comparisons. As the QIMME E01JH0011 dataset did not contain any PRE-specific or PRE-dominant features is was not included in this analysis