Occupants had a significant effect on total particle concentrations in the chamber air. Table 3 presents size-specific indoor/outdoor total particle number concentration ratios under three categories of occupancy conditions: zero, low , and high . In general, I/O ratios are influenced by particle removal through filtration of supply air, particle loss via deposition on chamber surfaces, and particle generation by occupants. An increase in filtration efficiency and/or deposition loss rate with particle size is evident in the data: the zero-occupancy I/O ratio declines from about 0.7 for 0.3–0.5 μm particles to 0.17–0.19 for particles larger than 5 μm. Table 3 shows that, for particles smaller than 2 μm, I/O ratios were not clearly affected by occupancy. Conversely, there was a clear impact of occupancy on larger particles that was stronger when the floor was exposed. Even for the particles most strongly impacted by occupancy , indoor particle concentrations with no or low occupancy loads correlate with outdoor particle concentrations, and the high occupancy load periods show elevated particle concentrations relative to this baseline level . The fraction of indoor particles larger than 5 μm that was attributable to occupancy was in the range 62–96% when the floor was covered and 84–87% when the floor was exposed . Taken together, the low I/O ratios and the attribution of particles to sources suggest a discernible yet relatively low total particle emission rate associated with occupancy in these experiments.There are several aspects of the way this chamber was operated that could contribute to a strong outdoor source relative to indoor emissions. For one,hydroponic bucket the air exchange rate was relatively high, 2.8 per hour, which is in the upper range of values measured in US office buildings and high compared to the central tendency in residences.
Given the size of the chamber, at its highest occupancy , building standards would require a lower minimum air exchange rate of ~ 1.3 per hour. Air was treated through MERV 7 filters, which offer only a modest level of filtration efficiency, although this filter quality is common for use in commercial buildings.Consistent with conditions that predict a strong outdoor source in the chamber, we observed a lower value for the percentage of human-associated taxa relative to other indoor microbiome studies. In the heavily occupied setting of the New York City subway system, approximately 20% of observed taxa were those associated with human skin. In settings more similar to our office environment than public transportation systems, Hospodsky et al. and Qian et al. reported that five human-associated bacteria taxa comprised 17% of all bacteria in their university classroom air samples while, in a separate university classroom study, Meadow et al. found that on average 8% of the sequences were human associated. In our study, the value was 4%. Accordingly, the proportion of sequences that were attributable to known human commensals in indoor air relative to outdoor air was enriched by a factor of 1.3 in our study, while in another study in a classroom the enrichment factor was 3.5. Despite the relatively smaller signal of human body-associated taxa we found in our study, the qualitative results—of a strong microbial link between outdoor and indoor air—matches other current studies that have paired outdoor and indoor air samples. Even in the presence of a higher occupancy load, the composition of bacteria in the air of the New York City subway system was most similar to outdoor air. A recent investigation in the Hong Kong subway found that the subway air and outdoor air were compositionally indistinguishable. Similarly, Meadow et al. showed that in university classrooms in the Pacific Northwest US, indoor air largely tracked outdoor air but with time lags that depended on the type of ventilation system.
This dominance of outdoor air relative to indoor occupancy as a source for bioaerosols is noteworthy, particularly relative to other studies that have compared unoccupied and occupied periods and other studies that have sampled bacteria on indoor surfaces. While we did observe microbial taxa associated with the human body cavity, the most striking effect of occupancy that we observed was of occupants as potential passive transport vectors of microbes from outdoors or from other indoor locations. This human-mediated dispersal was identified by the unique signal of the puffball Batterea being abundant in particular experiments when one specific person was present and absent in outdoor samples. Recently it was demonstrated that clothing could act as a secondary source of particles that were previously deposited on the fabric. Resident behavior as vectors transporting fungal material into homes has been observed in association with food , fuel wood chips, and on clothing. For example, Pasanen et al. report the movement of Acremonium, Alternaria, Botrytis, and Chryosporium from cowsheds into homes on the clothes of residents. And, more generally, research shows that soil resuspension and track-in can be major contributors of pollutants in house dust . Although we observed this human-mediated transport with Battarea only, it is likely important for the transport of other—perhaps most—outdoor microbes. And, in fact, integrating results from recent studies showing that occupants have a marked effect on the mass of airborne particles but not necessarily on the composition suggests that occupants may predominately be secondary emitters of environmental microbes that are largely sourced from outdoor air. That is, the microbial composition on occupants clothing may become similar to that of the outdoor air of the surrounding area, and thus the signature of occupants on indoor air microbial composition would be obscured as originating from the outdoor supply air. In this conceptualization, an occupied room would show quantitatively higher bioaerosol levels than an unoccupied room but the composition might be similar between the two cases, i.e. substantially independent of occupancy level.
We note that the potential for resuspension of environmental microbes rather than human associated microbes might be greater in the office-like chamber used in this study than in other settings that are consistently more heavily occupied. As a testing environment, this chamberhas relatively long periods of vacancy, especially compared to classrooms or homes. The microbial signature that would accumulate on interior surfaces owing to supply by ventilation and deposition by settling could come to resemble the composition of outdoor air. For those settings that are more heavily and regularly used, the settled dust that might later be resuspended could have a stronger human signature. Resuspension in the two cases—one with a history of high occupancy and another with low occupancy loads—would provide different occupancy associated fingerprints because of differences in the reservoir makeup. As our efforts at quantitative measures were hindered ,stackable planters further studies will be needed to more deeply explore the microbiological nature of occupancy-associated emissions. The single most abundant bacterial taxon in the floor dust, Bifidobacterium adolescentis, is a known human commensal, although it was rare in both the indoor and outdoor air samples. Curiously, we did not identify DNA from the fungus Malassezia, a recognized abundant human commensal. While Malassezia is highly abundant on skin, its relative abundance in environmental surveys of indoor environments varies.We did detect species and members of other genera that are known to contain human commensals, typically acting as yeasts on skin surfaces: Trichosporon, Rhodotorula mucilaginosa, Candida, and Cryptococcus. The simultaneous sampling of outdoor air aided in the interpretation of our results, as it showed that taxa that were observed indoors were not necessarily sourced indoors. Interestingly, skin-associated bacterial taxa were also observed in outdoor air, a pattern also reported by Qian et al., suggesting that even the outdoor air might have a human commensal signal in densely populated urban settings. Plus, there were five taxa identified as the fungus Wallemia, and two of those as Wallemia sebi specifically. Wallemia is typically regarded as an indoor contaminant whose presence indoors indicates growth internal to the structure. Only one of these Wallemia taxa, identified as Wallemiales sp., was not observed outdoors, and it was observed in only one indoor sampling period. The occurrence outdoors indicates that occurrence indoors should not be automatically attributed to indoor growth.In two of our sampling efforts, biomass yields were insufficient to undertake specific analyses. The first was quantitative PCR, which we employed to determine the fungal and bacterial biomass for each sampling period. Many samples were near detection limits.
Across the replicates for each sample, precision was low; that is, the variation of duplicated measurements of the same sample tended to be high. We present details on methods and results in S2 File. Quantification of biological material is an important vehicle for linking microbial composition with particle dynamics. With our sampling methods, we conclude that biomass yields were too low for robust conclusions based on qPCR. High-throughput sequencing is a more sensitive measure than qPCR and has been successfully reported for volumes of air similar to the ones we use here. The second detection limit challenge we encountered was with particle collection using the National Institute for Occupational Safety and Health 2-stage cyclone aerosol sampler, the BC 251. The first stage of the sampler fitted a disposable 15-mL Falcon collection tube and had a lower particle size cutoff of 3.7 μm aerodynamic diameter. The second stage employed a disposable 1.5-mL Eppendorf tube and collected particles in the size range 0.74– 3.7 μm. Attempts at amplification, after extraction following the protocol detailed above, were unsuccessful from the smaller size fraction and were inconsistent from the larger size fraction. Due to these patchy results, the cyclone samples were not processed further. The results suggested that in our study environment, the volume of air sampled was insufficient to obtain a signal clearly above detection limits. The NIOSH cyclone sampler requires a lower flow rate vacuum pump, and successful use of this sampler in environmental settings has been based on longer sampling times than we used here. Both qPCR and the NIOSH cyclone sampler rely on the extraction of DNA as the material for analysis. Although we used an extraction protocol developed for indoor environments that mirrors another laboratory’s practices, there may be modifications in the extraction process that could increase efficiency.Determining microbial composition with high-throughput sequencing is a powerful technique, and its utilization in indoor air research has great potential to offer important insight into bioaerosol dynamics. However, there are several steps in the processing pipeline that require particular attention, and best practices have not been agreed upon . One particular issue that we highlight here is treatment of low-abundance reads, sequences that are present at sparse levels. While some of these reads may represent true biological diversity, many may be spurious OTUs due to chimeric sequences, errors during PCR amplification, or imprecise OTU binning. As a quality assurance measure for our samples, we ran a mock community, which indicated that reads with fewer than 10 sequences in a sample were potentially spurious. Of the 84 taxa excluded from the mock community by this threshold, 10% were typed to the same taxonomy as the high-abundance reads, which would indicate spurious OTUs attributable to PCR error or imprecise OTU binning. The remaining low-abundance reads from the mock community were present at high abundance in other samples, which is indicative of barcode tag switching events . Results we report here were robust to inclusion or exclusion of these low abundance reads, supporting the general idea that these high-throughput methods are strong for questions that center on relationships among samples but can require more caution for exploring specific questions about community membership .The dire effects of climate change have plagued the livelihoods of rural communities in East Africa for generations. Land degradation attributed to human, drought and climate factors is increasingly threatening the four main agricultural communities: pastorals, semi-pastorals, agropastorals and mixed-farming, especially in the drylands regions of East Africa . Adaptation strategies, such as livestock mobility, diversification, feed purchases and animal restocking, have increasingly become unable to support their livelihoods . Moreover, dependence on livestock rearing is increasingly constrained by population growth, which results in the occupation of grazing areas by human settlements and urbanization . Although population growth and the gradual emergence of peri-urban centres are potential sources of market opportunities for livestock producers , the future livelihoods of rural communities given the contemporary climatic change have remained among one of the biggest challenges in the region .