Abstract
The “One Health” framework emphasizes the ecological relationships between soil, plant, animal and human health. Microbiomes play important roles in these relationships, as they modify the health and performance of the different compartments and influence the transfer of energy, matter and chemicals between them. Standardized methods to characterize microbiomes along food chains are, however, currently lacking. To address this methodological gap, we evaluated the performance of DNA extractions kits and commonly recommended primer pairs targeting different hypervariable regions (V3-V4, V4, V5-V6, V5-V6-V7) of the 16S rRNA gene, on microbiome samples along a model food chain, including soils, maize roots, cattle rumen, and cattle and human faeces. We also included faeces from gnotobiotic mice colonized with defined bacterial taxa and mock communities to confirm the robustness of our molecular and bioinformatic approaches on these defined low microbial diversity samples. Based on Amplicon Sequence Variants, the primer pair 515F-806R led to the highest estimates of species richness and diversity in all sample types and offered maximum diversity coverage of reference databases in in silico primer analysis. The influence of the DNA extraction kits was negligible compared to the influence of the choice of primer pairs. Comparing microbiomes using 515F-806R revealed that soil and root samples have the highest estimates of species richness and inter-sample variation. Species richness decreased gradually along the food chain, with the lowest richness observed in human faeces. Primer pair choice directly influenced the estimation of community changes (beta diversity) within and across compartments and may give rise to preferential detection of specific taxa. This work demonstrates why a standardized approach is necessary to analyse microbiomes within and between source compartments along food chains in the context of the One Health framework.
