Abstract
BACKGROUND: The indoor microbiome significantly impacts human health. Different sampling methods are used to characterize this environment, but it is unclear how these methods affect the resulting microbial profiles and health-related interpretations. This study aimed to compare the bacterial communities captured by two common methods, i.e., passive collection of settled air dust and active vacuuming of surface dust, and to evaluate how sampling choice influences epidemiological associations with respiratory health. METHODS: We collected paired settled air dust (n = 86) and vacuumed surface dust (n = 83) samples from 87 university dormitory rooms. The bacterial composition was characterized by sequencing the V3-V4 region of the 16S rRNA gene. We analyzed differences in bacterial diversity, taxonomic composition, predicted functional profiles, and associations with self-reported student health data (rhinitis, asthma, and respiratory infections). RESULTS: The two sampling methods captured drastically different bacterial communities (PERMANOVA R² = 0.65, p < 0.001). Vacuumed dust samples were dominated by the genus Pseudomonas (75.1% mean relative abundance), whereas air dust samples were more diverse and primarily composed of Ralstonia (15.6%), Pelomonas (11.3%), and Anoxybacillus (9.3%). Air dust samples exhibited significantly higher alpha diversity (observed taxa: 906.2 ± 179.6) compared to vacuumed dust (345.1 ± 82.3, p < 0.001). These compositional differences led to distinct predicted functional pathways and divergent associations with health outcomes. For asthma, airborne Collinsella was associated with strongly increased odds (OR 2.17, p = 0.003), while Paracoccus was associated with decreased odds (OR 0.062, p = 0.006). By contrast, associations in vacuumed dust were limited to taxa with small effect sizes, such as Peptoclostridium (OR 1.07, p = 0.004). Furthermore, while airborne genera like Slackia were associated with respiratory infections (OR 5.94, p = 0.005), no significant bacterial associations for respiratory infections were found in vacuumed dust. CONCLUSION: The choice between sampling settled air dust versus vacuumed surface dust provides profoundly different views of the indoor bacteriome, which can lead to contradictory conclusions in epidemiological studies of asthma and respiratory illness. Our findings underscore that the selection of an environmental sampling strategy is a critical methodological decision that can dictate the outcome and potential health implications of indoor microbiome research. Researchers must align their sampling method with a clear exposure hypothesis to ensure valid health risk assessments. Future indoor air quality standards and epidemiological studies should consider implementing multi-method sampling to capture a comprehensive profile of microbial exposure.