Importance of sample pre-treatments for the DNA-based characterization of microbiomes in cropland and forest soils

Soil-extracted DNA is widely used to characterise the composition of microbiomes. However, there is no commonly accepted protocol or understanding of how sample preservation and laboratory pre-incubation affect the results obtained. Here, we evaluated six different pre-treatments of sieved cropland and forest topsoils from five sites in Northern Germany, including freezing, drying, fresh analyses, and pre-incubations for 1 or 14 day prior to DNA extraction. Based on PCR-amplified 16S rRNA genes and ITS sequences, we characterized the abundance and diversity of bacteria, archaea and fungi, and compared it to a reference community composition obtained from samples immediately frozen with liquid nitrogen and without further manipulation. Irrespective of the pre-treatments, soil microbiomes could be differentiated according to their land use and sampling site. For prokaryotes, Acidobacteriota, Actinobacteriota or Thermoproteota, were more affected by pre-treatments than Bacteriodota, Firmicutes or Planctomycetota. For fungi, Basidiomycota and Mucoromycota, which predominated in forest soils, were more susceptible than the cropland predominant Ascomycota, indicating that different pre-treatments have a stronger impact on assessing the composition of fungal microbiomes from forest than from cropland. Differences between the reference and pre-treatments suggest that generally soil sieving enhanced the DNA extraction efficiency, except for dried soils, but it shifted the proportionality of detected taxa, as shown by reduced Shannon diversity. Overall, oven-drying had a more detrimental effect than air drying, freezing treatments or fresh analyses. For large scale monitoring projects, where fresh analysis is not possible, the gentle freezing approach as applied in this study would be the best choice. •DNA of six pre-treatments from forest and cropland soils analysed.•Irrespective of pre-treatments, microbiomes could be differentiated by site and land use.•Soil sieving appeared to increase DNA yield of fresh and frozen, but not of dried samples.•Compared to cropland forest fungal microbiomes are more pre-treatment sensitive.•Fresh analyses and gentle freezing provided the least biased results.