COMPUTATIONAL RESEARCH DAY

Building design and operation impact the accumulation and survival of microorganisms in indoor spaces. These microorganisms, together referred to as the indoor microbiome, have implications for human health and well being. However, the relative importance of factors like architecture compared to, e.g., human occupancy, remains unclear. This study aimed to identify putative sources of microbes found in dust and determine building characteristics that most strongly correlate with the human-derived content. Dust samples were collected from gyms, hallways, and offices in 40 athletic facilities (n=122 samples total) and processed with metagenomic sequencing to characterize microbial community structure. SourceTracker, a Bayesian-based algorithm, was applied to estimate the proportions of microbiota derived from human sources (i.e., skin, oral, vaginal, gut) using training data from the Human Microbiome Project. Relationships between putative human-sourced proportions of microbiota and building properties (e.g., space size, ventilation, occupancy, cleaning products/frequency, dust antimicrobial concentrations) were evaluated with beta regression. The dust microbiomes consisted of highly variable mixtures of few dominant and many low-abundance taxa likely deposited from humans (primarily skin; e.g., Propionibacterium, Staphylococcus, Corynebacterium) and the environment (e.g., Pseudomonas, Massilia, Micrococcus, Wolbachia). Despite the heterogeneity across samples, there was a strong negative relationship between proportions human-sourced microbes and the presence of operable windows or exterior doors. There were also weak, yet significant correlations between the human-derived dust microbes and consumer product-associated antimicrobials in dust (e.g., triclosan, parabens), human occupancy, and chemical cleaning frequency in gyms. Taken together, these data suggest that the formation of the indoor microbiome is a dynamic process involving the passive transport of microbes and potential interaction with chemicals from human skin. This process is dominated by dispersal, though potential impacts of cleaning and personal hygiene products may increase in spaces with limited windows and exterior doors, especially as spaces become more crowded.