Long-term storage of feces at -80 °C versus -20 °C is negligible for 16S rRNA amplicon profiling of the equine bacterial microbiome

Abstract

The development of next-generation sequencing technologies has spurred a surge of research on bacterial microbiome diversity and function. But despite the rapid growth of the field, many uncertainties remain regarding the impact of differing methodologies on downstream results. Sample storage temperature is conventionally thought to be among the most important factors for ensuring reproducibility across marker gene studies, but to date much of the research on this topic has focused on short-term storage in the context of clinical applications. Consequently, it has remained unclear if storage at -80 °C, widely viewed as the gold standard for long-term archival of feces, is truly required for maintaining sample integrity in amplicon-based studies. A better understanding of the impacts of long-term storage conditions is important given the substantial cost and limited availability of ultra-low temperature freezers. To this end, we compared bacterial microbiome profiles inferred from 16S V3-V4 amplicon sequencing for paired fecal samples obtained from a feral horse population from Sable Island, Nova Scotia, Canada, stored at either -80 °C or -20 °C for 4 years. We found that storage temperature did not significantly affect alpha diversity measures, including amplicon sequence variant (ASV) richness and evenness, and abundance of rare sequence variants, nor presence/absence, relative abundances and phylogenetic diversity weighted measures of beta diversity. These results indicate that storage of equine feces at -20 °C for periods ranging from a few months to a few years is equivalent to storage at -80 °C for amplicon-based microbiome studies, adding to accumulating evidence indicating that standard domestic freezers are both economical and effective for microbiome research.

Keywords: 16S; Amplicon; DNA metabarcoding; Equine; Equus ferus caballus; Horse; Long-term storage; Microbiome; Protocol; Sequencing.

Figure 1. Relative abundance of bacterial families identified in fecal samples collected from feral horses on Sable Island, Nova Scotia, Canada, using 16S rRNA sequencing.

Samples from eight horses were divided into aliquots, stored at both −20 °C and −80 °C for 4 years, and characterized using 16S amplicon sequencing. Taxonomy was assigned to amplicon sequence variants (ASVs), and abundances were converted to proportional values then aggregated to the family level. Families with low abundances within a sample (<1%) were pooled for clarity.

Figure 2. Alpha diversity (richness (A), Choa1 (B) and Shannon indices (C)) of 16S V3–V4 Amplicon Sequence Variants (ASV) for paired equine fecal samples (aliquots) stored at −20 °C and −80 °C for 4 years.

Samples were collected from separate individuals on Sable Island, Nova Scotia, Canada. Paired samples (aliquots) are denoted by color.

Figure 3. Non-metric multidimensional scaling (NMDS) of equine gut bacterial communities inferred from fecal samples stored at −20 °C and −80 °C.

Ordinations were conducted on (A) Euclidean (stress = 0.08), (B) Jaccard (stress = 0.10), (C) Unweighted UniFrac (stress = 0.10) and (D) weighted UniFrac (stress = 0.14) distance matrices from Amplicon Sequencing Variant (ASV) count data. Paired samples share the same color and temperature is denoted by circles (−20 °C) and triangles (−80 °C). Red and blue ellipses represent 95% confidence intervals using the t-distributions for samples stored at −20 °C and −80 °C, respectively.