306. Short-term Provision of Moderate Dietary Starch Induces Compositional and Functional Shifts in the Fecal Microbiome Independent of Alpha Diversity Metrics in Yearling Quarter Horses.

Abstract Gastrointestinal microbial communities are dynamic and adapt to host diet, environment, and physiological status. However, fecal metabolites are rarely quantified alongside species diversity and abundance in microbiome research, limiting functional interpretations of microbial shifts. To test the hypothesis that a starch-based diet would result in microbial species and functional metabolite alterations compared to a fiber-based diet, Quarter Horses (mean±SD 16±1mo; 337±30kg) were acclimated to a basal diet of 1% bodyweight (BW)/d of a high-fiber concentrate (36.3% NDF, 8.5% starch) and 1.5% BW/d coastal Bermudagrass hay in individual runs. Following 21d, horses were randomly assigned to receive either the basal diet (CON; 7 fillies, 8 geldings) or hay plus 1% BW/d of an isocaloric, isonitrogenous starch concentrate (STARCH; 24.1% NDF, 28.7% starch; 8 fillies, 7 geldings) for 24d. Fecal metagenomics, primary metabolites and lipidomics were evaluated on d0 and 21. Fecal metagenomic data were quality-filtered, and metagenomic reads were classified and analyzed in R using diversity metrics, PERMANOVA, LEfSe, and regression approaches to assess treatment effects. Fecal metabolites were extracted via biphasic separation and analyzed using LC-HRMS/MS for lipidomics and GC-TOFMS for primary metabolites, followed by peak detection, alignment, normalization, and pathway annotation in Metaboanalyst. Alpha diversity, measured by Shannon and Simpson indices, did not differ by treatment or time, with microbial community richness remaining stable through d21 (Kruskal-Wallis, Shannon ∼ TP, P = 0.51). Despite this, beta diversity revealed significant shifts in microbial community structure. Principal Coordinate Analysis (PCoA) based on Bray-Curtis dissimilarity indicated that the composition was significantly influenced by the interaction of treatment and time (PERMANOVA, F = 2.20, R²=0.07, P = 0.003). Pairwise comparisons revealed the nature of shifts: CON horses showed minor changes from d0 to 21 (P = 0.05), whereas STARCH-treated horses exhibited a highly significant shift in community structure (P = 0.003). Consequently, STARCH and CON horses had distinct microbial communities at d21 (P = 0.04), suggesting that diet drove microbiome successional trajectory. Specifically, CON horses had greater Lactobacillaceae (∼5.7% vs. ∼2.4% in STARCH), while STARCH had greater Lachnospiraceae (∼38% vs. ∼32% in CON). Correspondingly, fecal metabolomics revealed that 18 fecal lipid metabolites were significantly altered between CON and STARCH at d21 (P < 0.05). STARCH horses demonstrated positive fold changes in quinoline derivatives, acylcarnitine, and phosphorylcholine metabolite clustering (P = 0.04) while metabolites relating to ceramide synthesis clustered tightly in CON horses (P ≤ 0.05). Primary metabolite profiling revealed 18 additional differentially expressed fecal metabolites between CON and STARCH horses at d21, supporting functional divergence. Notably, 2,8-dihydroxyquinoline, indicative of microbial metabolic flux, was elevated in STARCH compared to CON horses, demonstrating the greatest log fold change between metabolites (logFC=1.03, P < 0.0001). Collectively, these data demonstrate that while alpha diversity remained unchanged, metagenomic and metabolomic analyses reveal clear compositional and functional divergence, highlighting that diversity metrics alone may underestimate biologically meaningful microbiome alterations