Impact of sequencing approach on ecological inference from highly variable bat gut microbiomes

Understanding the functional potential of animal gut microbiomes requires analytical approaches that can accurately capture both taxonomic and functional diversity, particularly in hosts with samples with low and variable microbial DNA fractions, such as bats. Here, we compared 16S amplicon sequencing and genome-resolved metagenomics for characterising the gut microbiomes of three sympatric insectivorous bats (Pipistrellus kuhlii, Hypsugo ariel, and Cnephaeus bottae). Amplicon sequencing recovered 3,536 amplicon sequence variants (ASVs) spanning 29 microbial phyla, including five archaeal groups, while metagenomics yielded 135 metagenome-assembled genomes (MAGs) across 15 bacterial phyla. Alpha and beta diversity patterns differed significantly between approaches. Applying prevalence- and abundance-based filtering to amplicon data improved alpha diversity congruence with metagenomic profiles, indicating that low-abundance and rare ASVs inflate diversity estimates, while the low microbial DNA fractions in bat faeces challenge genome reconstruction of such taxa. Functional reconstructions diverged notably, with metagenomics-based functional profiles yielding consistently lower metabolic capacities than those indirectly inferred from 16S amplicon sequencing, both before and after prevalence- and abundance-based filtering, due to incompleteness of many reconstructed genomes. However, metagenome-assembled genomes harboured metabolic capacities missed by the indirect functional inference of metabarcoding, which largely rely on genomes of strains characterised in model organisms. Together, these findings highlight the influence of chosen methodology, host species, and data filtering on microbiome profiling in taxa with low microbial biomass. We propose practical guidelines to improve analytical consistency and reliability when studying gut microbiomes of bats and other hosts with similarly variable microbial communities.