Fig. 5

Microbial community dynamics in (non-)extractive chain elongation continuous reactors. Initial substrates were l-lactate and acetate. A Shannon diversity index of inoculum, R1 and R2 microbiomes. B Shannon diversity index of suspended microbiomes from non-extractive (5-N) and extractive chain elongation at pH 5.0 (5-E) and pH 7.0 (7-E); and biofilm microbiomes from extractive chain elongation at pH 5.0 (5-EB) and pH 7.0 (7-EB). Boxplots show the interquartile range (IQR) in boxes divided by median values (horizontal lines) with whiskers depicting ± 1.5 IQR (A, B). C Distance-based Redundancy Analysis (dbRDA) using Bray–Curtis dissimilarity index; ASVs relative abundance as response variables; and environmental parameters as explanatory variables (constraints). Environmental parameters considered were: reactor (R1, R2), pH, sunflower oil (SO) (presence/absence) and steady-state electron selectivities (propionate, nC4, nC6, nC7, nC8, MCC and H2; Additional file 1: Tables S2 and S3). Concentration ellipses depict confidence intervals with α = 0.05. Significance code: ‘***’ associated with a variable at P < 0.0005; ‘*’ associated with a variable at P < 0.01; and ‘.’ associated with a variable at P < 0.05. D Canonical Correspondence Analysis (CCA) using Hellinger transformation and same constraints as in dbRDA. CCA triplot shows top 11 ASVs scaled proportionally to eigenvalues. Collinear (redundant) environmental parameters (MCC and H₂) were automatically dropped out in both dbRDA and CCA (C, D). E, F Differential abundance analyses at ASV level of microbiomes developed at different pH conditions (E) and between reactors (F). Differential abundance at different pH was analysed using all samples from both reactors. II-a and II-b biofilms samples were left out when comparing microbiomes between reactors. E, F. Analyses were done using CSS-normalized 16S rRNA gene sequencing data of ASVs with > 0.01% of total counts