Fig. 4

Baseline taxonomic predictors of microbiota departure from baseline. Volcano plot showing coefficients from penalized generalized estimating equations (PGEE) and their corresponding p values. The model included clr abundances of baseline OTUs collapsed at genus level, decay averaged antibiotic histories, age, sex, body mass index, use of parenteral nutrition, baseline Shannon index, sample depth, and interval from baseline as potential predictors of the Aitchison distance between non-baseline microbiota and baseline (pre-treatment) microbiota for the same patient. Patient ID was a random effect. The tuning parameter was 0.21, determined using fourfold cross validation. A first-order autoregressive correlation structure was assumed. The horizontal dashed line defines statistical significance (p < 0.05). Circles colored red or blue represent significant taxa with a coefficient > 10⁻³ (red) or < − 10⁻³ (blue), respectively. These taxa at baseline were associated with the extent of departure from baseline in subsequent samples and their corresponding circles were magnified for better visualization