Comparison of oral collection methods for 16S rRNA gene and shotgun metagenomic sequencing

To understand how sample collection affects oral microbiome studies, we evaluated the comparability of unpreserved saliva, saliva in glycerol, and mouthwash samples, their room temperature stability, and intraindividual stability over 6 months. Saliva and mouthwash samples were collected from 20 healthy participants 6 months apart. Saliva was divided, with half preserved in glycerol. Some aliquots were frozen immediately, while others were stored at room temperature for a week. DNA was extracted using the PowerSoil Pro and 16S rRNA gene, and shotgun metagenomic sequencing was conducted. Intraclass correlation coefficients (ICCs) from taxonomic and functional tables were compared to assess variability. We estimated sample size requirements based on the intraindividual stability over 6 months. Saliva in glycerol appeared more similar to unpreserved saliva than mouthwash, with higher median ICCs at genus (0.88 vs 0.60), species (0.92 vs 0.64), and gene levels (0.84 vs 0.36; all P < 0.01). Room temperature storage affected saliva in glycerol more than mouthwash (median genus-level ICC = 0.65). No significant differences were observed at the gene level. Intraindividual stability over 6 months was moderate. To detect an odds ratio of 1.5 with one sample per individual, estimated sample sizes ranged from 665 (common species) to 219,547 (rare species). Oral microbiome stability varies by collection method; mouthwash provides greater room temperature stability and may be preferable when immediate freezing is not feasible. For epidemiological studies, consistent use of a single collection method and inclusion of longitudinal sampling can improve reproducibility and power to detect associations with health outcomes. IMPORTANCE The oral microbiome plays a key role in health and disease, yet methodological inconsistencies in sample collection and processing can introduce variability and limit comparability across studies. This study investigates the impact of different oral sample collection methods on microbiome profiling and their stability over time. We demonstrate that sample type significantly influences taxonomic and functional microbiome profiles, with mouthwash showing greater stability during delayed processing and saliva in glycerol more closely resembling fresh saliva. Importantly, intraindividual microbial communities were only moderately stable over 6 months, emphasizing the need for consistent sampling protocols and consideration of temporal variation. These findings have direct implications for microbiome study design, highlighting that methodological choices can affect reproducibility, statistical power, and biological interpretation. Our results support the use of mouthwash as a practical alternative when freezing is delayed and underscore the value of longitudinal sampling for detecting biologically meaningful changes.