Evaluating stool microbiome integrity after domestic freezer storage using whole-metagenome sequencing, genome assembly, and antimicrobial resistance gene analysis

Abstract

The gut microbiome is crucial for host health. Early childhood is a critical period for the development of a healthy gut microbiome, but it is particularly sensitive to external influences. Recent research has focused on using advanced techniques like shotgun metagenome sequencing to identify key microbial signatures and disruptions linked to disease. For accurate microbiome analysis, samples need to be collected and stored under specific conditions to preserve microbial integrity and composition, with -80°C storage considered the gold standard for stabilization. This study investigates the effect of domestic freezer storage on the microbial composition of stool samples obtained from 20 children under 4 years of age with the use of shotgun metagenome sequencing. Fresh stool samples were aliquoted into sterile tubes, with one aliquot stored at 4°C and analyzed within 24 hours, while others were frozen in domestic freezers (below -18°C) and analyzed after 1 week, 2 months, and 6 months. Assessments of contig assembly quality, microbial diversity, and antimicrobial resistance genes revealed no significant degradation or variation in microbial composition.

Importance: Most prior studies on sample storage have relied on amplicon sequencing, which is less applicable to metagenome sequencing-given considerations of contig quality and functional gene detection-and less reliable in representing microbial composition. Moreover, the effects of domestic freezer storage for at-home stool collection on microbiome profiles, contig quality, and antimicrobial resistance gene profiles have not been previously investigated. Our findings suggest that stool samples stored in domestic freezers for up to 6 months maintain the integrity of metagenomic data. These findings indicate that domestic freezer storage does not compromise the integrity or reproducibility of metagenomic data, offering a reliable and accessible alternative for temporary sample storage. This approach enhances the feasibility of large-scale at-home stool collection and citizen science projects, even those focused on the more easily perturbed early life microbiome. This advancement enables more inclusive research into the gut microbiome, enhancing our understanding of its role in human health.

Keywords: fecal; gut; infant; microbiota; shotgun sequencing; −20°C.

Fig 1

(a) PCoA based on the Aitchison distance matrix, illustrating the beta diversity of the microbial communities. Each dot represents a sample, colored by child ID (S01–S20). (b) Mean Aitchison distance between samples at different time points, representing compositional differences. The mean distance is calculated as the average pairwise Aitchison distance for all samples within each time point, with individual dots representing the pairwise distances between samples. (c) Alpha diversity metrics across time points: the Shannon diversity index (top panel) and observed features (bottom panel) are shown at 0 weeks (0W), 1 week (1W), 2 months (2M), and 6 months (6M). The Shannon index reflects diversity within samples, while observed features indicate species richness. Thick black lines represent the mean values, with shaded gray areas denoting standard deviation. The Wilcoxon rank-sum test with Bonferroni correction was used to assess differences, with no significant differences (ns) observed between the time points for panels b and c.

Fig 2

(a) Stacked area plots showing the relative abundance of the top 20 most abundant species across different time points (0W = week 0, 1W = week 1, 2M = month 2, and 6M = month 6) for each child (S01–S20). Each color represents a different species, illustrating the dynamic changes and stability in the microbial community composition over the first 6 months of life. (b) Line plots depicting the relative abundance trends of the most abundant genera. Each plot shows the relative abundance over time per individual child data (thin lines) and the overall mean trend (thick line).

Fig 3

(a) Principal coordinate analysis (PCoA) plot depicting beta diversity based on the Jaccard distance matrix, calculated from antimicrobial resistance (AMR)-conferring genes. Each point represents a sample, colored by child (S01–S20), as indicated in the legend, with the percentage of variance explained by PC1 and PC2. (b) Heatmap of the top 20 most abundant ARO terms, displaying log-transformed normalized reads per million (RPM) values. (c) Line plots depicting the relative abundance trends of the most abundant AMR-conferring genes. Each plot shows RPM over time per individual child data (thin lines) and the overall mean trend (thick line). The y-axis is displayed on a pseudo-log scale (plog, log(1 + x)) to enhance visualization of small or zero values. The Wilcoxon rank-sum test with Bonferroni correction was used to assess differences, with no significant differences (ns) observed between the time points (0W = week 0; 1W = week 1; 2M = month 2; 6M = month 6).

Fig 4

N50 (a) and L50 (b) metrics in stool microbiome samples stored over time (0W = week 0, 1W = week 1, 2M = month 2, and 6M = month 6). Line plots show individual child data (gray lines) and the overall mean trend (red line), with shaded red areas representing the standard deviation. The Wilcoxon rank-sum test with Bonferroni correction was used to compare metrics across time points, with no significant differences (ns) observed.