Antibiotic growth promoter and phytogenic feed additive consistently alter microbial community structure in chicken cecum

Background: Efforts to replace antibiotic growth promoters (AGPs) in livestock are often hindered by a limited mechanistic understanding of how sub-therapeutic antibiotic doses enhance animal growth. Since AGP concentrations are typically too low to directly suppress pathogens, their effects on the gut microbiome, particularly its ecological dynamics, warrant closer investigation. A critical but underexplored dimension is how these additives influence the structure and stability of microbial communities as interconnected ecosystems. Methods: , an alternative phytogenic feed additive (PFA) on cecal microbiome dynamics in broiler chickens. Using metagenomic data from a repeated cross-sectional randomized controlled trial of 96 broiler chickens assigned to three dietary groups: Basal (Control), AGP and PFA, we constructed microbial co-occurrence networks using Spearman's correlation for birds raised on basal, AGP-, or PFA-supplemented diets at key developmental stages (Day 3, 14, 21, and 35). We assessed changes in network topology, modular organization and node centrality. We evaluated whether the network-prioritized keystone taxa could discriminate among diets using a Random Forest classifier. Results: Compared to the Control group, both AGP and PFA treatments induced consistent shifts in network topology, including reduced connectivity, increased modularity, increased percentage of positive interactions, enhanced mucosa connectivity, and improved structural robustness over experiment time. Overall, these treatment-induced changes were more pronounced under AGP than under PFA. Despite these changes, we identified conserved subgraphs with stable interconnections across diets and time points during the experiment. The node centrality analysis revealed condition-specific keystone taxa, but Linear Discriminant Analysis (LDA) and Random Forest (RF) struggled to accurately differentiate between diets using their abundance, particularly between PFA and the two other groups. Conclusion: Our findings reveal that feed additives can reshape gut microbial dynamics without producing marked compositional shifts. The consistent network-level changes observed for both AGP and PFA highlight the value of ecological network analysis in uncovering microbial community responses. These insights improve our understanding of cecal microbiome responses in chickens, highlight potential modes of action of AGPs, and offer a comparative framework for assessing the microbial impacts of alternative feed additives.