DNA extraction method affects diversity indices of ecosystem microbiota

Abstract Background Metataxonomic studies of ecosystem microbiotas require the simultaneous processing of samples with contrasting physical and biochemical traits to guarantee that diversity indices are comparable. However, DNA extraction can induce undesired bias in the microbial community composition and hamper the detection of meaningful biological patterns in metataxonomic datasets. Therefore, to compare microbiotas of samples with different requirements for DNA extraction, the methodological variation resulting from the adoption of different extraction method needs to be considered and quantified. There are currently no studies in which different DNA extraction kits were compared across the main components of terrestrial ecosystems, such as soil, rhizosphere, invertebrate animals, and mammalian feces. Results Five DNA extraction kits designed for the extraction of nucleic acids from soil, feces and animal tissue were used to investigate the composition and diversity of microbota of a subset of samples typically studied in terrestrial ecosystems such as soil (bulk soil and rhizosphere soil), invertebrate taxa (Nematoda, Coleoptera) and mammalian feces ( B. taurus and Lepus spp.) Our data shows that each DNA extraction method leads to significant differences in the relative abundance of several prokaryotic taxa. Furthermore, we demonstrate that kit selection has the potential to bias both alpha and beta diversity estimates. Importantly, the impact of DNA extraction kit on sample diversity estimates varies according to the sample type, with mammal feces and soil samples showing the most and least consistent results across DNA extraction kits, respectively. Conclusion The selection of DNA extraction kits for large-scale microbial ecological studies involving the comparison of microbiotas of heterogeneous sample types should consider sample susceptibility to methodological variation. From the comparisons of five DNA extraction kits commonly used to study soil, feces and invertebrates, we showed that NucleoSpin® Soil (MACHEREY-NAGEL; MNS) and QIAamp® Fast DNA Stool Mini (QIAGEN; QST) were associated with the highest and lowest alpha and beta diversity estimates across the different sample types, respectively. These observations indicate that MNS efficiently captures both the composition and diversity of microbial communities found in the terrestrial samples considered in this study, and we recommend its use for any large-scale microbiota study of terrestrial ecosystems.