Genome-resolved metagenomics reveals lineage-specific phylogenomic and functional differentiation among...

Microbial carbon processing in mangrove and terrestrial forests remains poorly resolved at the level of individual bacterial genomes. Here, shotgun metagenomics was used to compare bacterial communities and genome-encoded carbon-processing potential across mangrove sediments, terrestrial forest soils, and associated litter from Goa, India. Genome reconstruction recovered 15 bacterial metagenome-assembled genomes (MAGs), including eight phylogenomically distinct, species-unresolved lineages identified through genome-based taxonomic and relatedness analyses. Community profiles broadly overlapped between habitats, with the strongest contrasts emerging among reconstructed lineages and their functional repertoires. The recovered genomes differed in carbohydrate-processing and bioelement-associated functions, revealing lineage-specific organization of carbon-processing potential. The mangrove-associated Paracoccus marcusii encoded RuBisCO-associated genes together with C1- and sulfur-associated functions, representing a distinctive carbon- and redox-associated genomic repertoire. Validated carbonic anhydrase classes also varied among bacterial lineages. Carbohydrate-active enzyme profiles revealed pronounced substrate-level differentiation, with the mangrove-derived Verrucomicrobiota-affiliated JAAUTS01 lineage showing the broadest divergence from its close reference across multiple plant- and detritus-associated carbohydrate functions. Terrestrial-derived Pseudomonas_E and Sphingobacteriaceae-affiliated lineages displayed distinct restructuring across substrate-, bioelement-, and carbonic-anhydrase-associated traits, whereas several other MAG-reference pairs remained comparatively conserved, showing that genomic divergence does not uniformly translate into broad functional novelty. Exploratory carbon-strategy indices showed directionally higher stabilization scores among mangrove-derived MAGs and provided a transparent framework for testing genome-derived carbon traits against independently replicated environmental and process measurements. These findings show that carbon-processing potential is organized primarily at the lineage and functional-repertoire level and identify candidate genomic traits for future microbiome-informed carbon monitoring across blue-carbon and terrestrial forest systems.