Decoding Forage-Driven Microbial–Metabolite Patterns: A Multi-Omics Comparison of Eight Tropical Silage Crops

Tropical forage crops vary widely in biochemical composition, resulting in inconsistent silage quality. Understanding how plant traits shape microbial and metabolic networks during ensiling is crucial for optimizing fermentation outcomes. Eight tropical forages—Sorghum bicolor (sweet sorghum), Sorghum × drummondii (sorghum–Sudangrass hybrid), Sorghum sudanense (Sudangrass), Pennisetum giganteum (giant Napier grass), Pennisetum purpureum cv. Purple (purple elephant grass), Pennisetum sinese (king grass), Leymus chinensis (sheep grass), and Zea mexicana (Mexican teosinte)—were ensiled under uniform conditions. Fermentation quality, bacterial and fungal communities (16S rRNA and ITS sequencing), and metabolite profiles (untargeted liquid chromatography–mass spectrometry, LC-MS) were analyzed after 60 days. Sweet sorghum and giant Napier grass showed optimal fermentation, with high lactic acid levels (111.2 g/kg and 99.4 g/kg, respectively), low NH4+-N (2.4 g/kg and 3.1 g/kg), and dominant Lactiplantibacillus plantarum. In contrast, sheep grass and Mexican teosinte exhibited poor fermentation, with high NH4+-N (6.7 and 6.1 g/kg) and Clostridium dominance. Fungal communities were dominated by Kazachstania humilis (>95%), while spoilage-associated genera such as Cladosporium, Fusarium, and Termitomyces proliferated in poorly fermented silages. Metabolomic analysis identified 15,827 features, with >3000 significantly differential metabolites between silages. Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment revealed divergence in flavonoid biosynthesis, lipid metabolism, and amino acid pathways. In the sweet sorghum vs. sheep grass comparison, oxidative stress markers ((±) 9-HODE, Agrimonolide) were elevated in sheep grass, while sweet sorghum accumulated antioxidants like Vitamin D3. Giant Napier grass exhibited higher levels of antimicrobial flavonoids (e.g., Apigenin) than king grass, despite both being dominated by lactic acid bacteria. Sorghum–Sudangrass hybrid silage showed enrichment of lignan and flavonoid derivatives, while Mexican teosinte accumulated hormone-like compounds (Gibberellin A53, Pterostilbene), suggesting microbial dysbiosis. These findings indicate that silage fermentation outcomes are primarily driven by forage-intrinsic traits. A “forage–microbiota–metabolite” framework was proposed to explain how plant-specific properties regulate microbial assembly and metabolic output. These insights can guide forage selection and development of precision inoculant for high-quality tropical silage.