Microbial biogeography of a university campus

Abstract

Background: Microorganisms are distributed on surfaces within homes, workplaces, and schools, with the potential to impact human health and disease. University campuses represent a unique opportunity to explore the distribution of microorganisms within built environments because of high human population densities, throughput, and variable building usage. For example, the main campus of the University of Waterloo spans four square kilometres, hosts over 40,000 individuals daily, and is comprised of a variety of buildings, including lecture halls, gyms, restaurants, residences, and a daycare.

Results: Representative left and right entrance door handles from each of the 65 buildings at the University of Waterloo were swabbed at three time points during an academic term in order to determine if microbial community assemblages coincided with building usage and whether these communities are stable temporally. Across all door handles, the dominant phyla were Proteobacteria, Firmicutes, Actinobacteria, and Bacteroidetes, which comprised 89.0 % of all reads. A total of 713 genera were observed, 16 of which constituted a minimum of 1 % of the 2,458,094 classified and rarefied reads. Archaea were found in low abundance (~0.03 %) but were present on 42.8 % of the door handles on 96 % of buildings across all time points, indicating that they are ubiquitous at very low levels on door handle surfaces. Although inter-handle variability was high, several individual building entrances harbored distinct microbial communities that were consistent over time. The presence of visible environmental debris on a subset of handles was associated with distinct microbial communities (beta diversity), increased richness (alpha diversity), and higher biomass (adenosine 5'-triphosphate; ATP).

Conclusions: This study demonstrates highly variable microbial communities associated with frequently contacted door handles on a university campus. Nonetheless, the data also revealed several building-specific and temporally stable bacterial and archaeal community patterns, with a potential impact of accumulated debris, a possible result of low human throughput, on detected microbial communities.

Fig. 1

Relative abundance of the 47 phyla found on 65 university campus building door handles. The Average data represent pooled sequences from all three time points

Fig. 2

The relative abundance of the most prevalent bacterial and archaeal OTUs. a The 16 bacterial OTUs that make up >1 % of all reads and the phylum to which they belong. b All 26 archaeal OTUs and their respective phyla

Fig. 3

Diversity of the 65 university campus buildings. a The average number of OTUs observed for each building. Each point consists of up to six samples, with two per sample date. Error bars represent the standard deviation. b Shannon index for each door, organized by building. Blue points represent swabs that were associated with visible debris, whereas red points indicate swabs without debris. The bottom row is color coded according to the faculty or specific campus function of the building

Fig. 4

PCoA plot calculated using the Bray-Curtis dissimilarity metric. The 383 samples from all time points are included and organized by building. Polygons encompass all samples from 25 buildings that remained relatively stable over the course of the study. Samples from remaining buildings, which were highly variable over time, are shown as gray points. Abbreviations are explained in Additional file 3: Table S2

Fig. 5

PCoA plot calculated using the Bray-Curtis dissimilarity metric. The 383 samples from all time points are included and organized by debris (a) and throughput (b). Inset contains the quantity of ATP detected on swabs from 10 university campus door handles that were covered in visible debris and 10 without debris