Impact of storage and extraction methods on peat soil microbiomes

Abstract

Recovered microbial community structure is known to be influenced by sample storage conditions and nucleic acid extraction methods, and the impact varies by sample type. Peat soils store a large portion of soil carbon and their microbiomes mediate climate feedbacks. Here, we tested three storage conditions and five extraction protocols on peat soils from three physicochemically distinct habitats in Stordalen Mire, Sweden, revealing significant methodological impacts on microbial (here, meaning bacteria and archaea) community structure. Initial preservation method impacted alpha but not beta diversity, with in-field storage in LifeGuard buffer yielding roughly two-thirds the richness of in-field flash-freezing or transport from the field on ice (all samples were stored at -80 °C after return from the field). Nucleic acid extraction method impacted both alpha and beta diversity; one method (the PowerSoil Total RNA Isolation kit with DNA Elution Accessory kit) diverged from the others (PowerMax Soil DNA Isolation kit-High Humic Acid Protocol, and three variations of a modified PowerMax Soil DNA/RNA isolation kit), capturing more diverse microbial taxa, with divergent community structures. Although habitat and sample depth still consistently dominated community variation, method-based biases in microbiome recovery for these climatologically-relevant soils are significant, and underscore the importance of methodological consistency for accurate inter-study comparisons, long-term monitoring, and consistent ecological interpretations.

Keywords: Extraction; Methods; Microbiome; Peatland; Soil; Storage; Stordalen.

Figure 1. Overview of Methods.

Soil from all three habitats, from two depths, were stored under three conditions and tested with a single extraction protocol (‘S’). Material stored in LifeGuard was extracted by five methods. Triplicate extractions were performed for each method combination. Extraction success and differential bias was characterized by DNA and RNA yield and quality, and by 16S rRNA amplicon sequencing. Sample Storage Experiment: 3 sites × 2 depths × 3 storage methods × 3 replicates × 1 extraction method. Nucleic Acids Extraction Experiment: 3 sites × 2 depths × 1 storage method × 3 replicates × 5 extraction methods. Abbreviations: Storage: LG, LifeGuard buffer; LN, liquid nitrogen; NAK, no buffer; Extraction: S, IsoGenie standard protocol; SL, IsoGenie Standard protocol with LifeGuard removed before extraction; SR, IsoGenie standard protocol with alternative reagents; PSH, PowerMax Soil DNA Isolation kit-High Humic Acid Protocol; RC, PowerSoil Total RNA Isolation kit with DNA elution accessory kit.

Figure 2. Impacts of storage conditions on microbial community structure and diversity.

(A) Principal coordinate analysis (PCoA) based on Bray-Curtis dissimilarities of microbiome profiles (via 16S rRNA amplicon sequencing) for the three storage methods. The R² values and associated p-values from an adonis2 test (PERMANOVA), indicating the proportion of variance explained by habitat, depth, and storage, are provided. Storage methods are indicated by shape: circles = LifeGuard (LG), triangles = liquid nitrogen (LN), and squares = on ice without any preservation buffer (NAK). The three habitats are indicated by color, brown = palsa, green = bog, and blue = fen. (B–C) Two alpha diversity metrics, Shannon’s diversity (B) and Richness as Observed ASVs (C) were compared across the three storage conditions. Statistical significance for alpha diversity metrics was evaluated by a Kruskal-Wallis non-parametric test between groups; pairwise comparison p-values are denoted above the associated lines, and the significance of overall storage condition impact on microbiome composition is indicated at the top of each panel.

Figure 3. Impacts of extraction protocols on microbial community structure and diversity.

(A, B, E, F) Principal coordinate analysis (PCoA) based on Bray-Curtis dissimilarities of microbiome profiles (via 16S rRNA amplicon sequencing) for five different extraction methods on samples stored in LifeGuard buffer. The R² values and associated p-values from an adonis2 test (PERMANOVA), indicating the proportion of variance explained by habitat and extraction are provided. (A–D) Abundance data calculated from ASVs. (E–H) Abundance data based on 97%-identity OTUs. Shallow samples (10–14 cm), Deep samples (30–34 cm). Extraction methods: S, IsoGenie standard protocol; SL, IsoGenie Standard protocol with LifeGuard removed before extraction; SR, IsoGenie standard protocol with alternative reagents; PSH, PowerMax Soil DNA Isolation kit–High Humic Acid Protocol; RC, PowerSoil Total RNA Isolation kit with DNA elution accessory kit. Two alpha diversity metrics, Shannon’s diversity (D, H) and observed richness (C, G) were compared across the five extraction protocols. Statistical significance for alpha diversity metrics was evaluated by a Wilcoxon rank sum tests between groups, adjusted via the Benjamini-Hochberg correction; significant p-values (p < 0.05) are denoted above the associated groups, where a significant difference between two extraction methods is indicated by a different lettering. Groups with the same letter display insignificant differences (p > 0.05). Common legend is applicable to both plots.

Figure 4. Recovery of differentially abundant lineages across extraction methods.

(A, C) Total differentially-recovered lineages for each of the five extraction methods, for shallow (A) and deep (C) samples. Stacked barcharts show the average relative abundance of differentially-recovered lineages (i.e., either over- or under-represented lineages), colored by phylum of bacteria or archaea. (B, D) Heatmap of the significantly differentially abundant ASVs (rows), with sample annotation for location/habitat and extraction method (columns), for shallow (B) and deep (D). Rows and columns are hierarchically clustered using the Ward.D2 method, and the data were normalized using log-transformed counts from DESeq2. Note slightly different scale for B and D. Extraction methods: S, IsoGenie standard protocol; SL, IsoGenie Standard protocol with LifeGuard removed before extraction; SR, IsoGenie standard protocol with alternative reagents; PSH, PowerMax Soil DNA Isolation kit–High Humic Acid Protocol; RC, PowerSoil Total RNA Isolation kit with DNA elution accessory kit.

Figure 5. Comparative analysis of microbial community distances based on different extraction methods.

Boxplots and violin plots illustrating the pairwise Bray-Curtis distances (A, C) and Weighted Unifrac distance (B, D) between microbial communities for various extraction methods for both ASVs (top row; A, B) and 97%-OTUs (bottom row; C, D). Each extraction method is compared against others within the same habitat and depth strata (n = 213 for all groups except RC, which n = 204). Significant pairwise comparisons between extraction methods are denoted with letters, indicating adjusted p-values post multiple testing correction (Wilcoxon rank-sum tests with Benjamini-Hochberg correction). The jittered points represent comparisons against a specific extraction method, colored according to the compared method. Note different y-axis scales. Extraction methods: S, IsoGenie standard protocol; SL, IsoGenie Standard protocol with LifeGuard removed before extraction; SR, IsoGenie standard protocol with alternative reagents; PSH, PowerMax Soil DNA Isolation kit–High Humic Acid Protocol; RC, PowerSoil Total RNA Isolation kit with DNA elution accessory kit.