Advancing understanding of how microbial communities are influenced by building characteristics and applying that understanding requires undertaking studies that integrate information from the microbial and building science fields. This chapter reviews a range of tools needed to study the characteristics of buildings and building-associated microbiomes. Topics discussed include standardized methods for collecting data on built environments and microbial communities and tools for analyzing these data to improve understanding of microbiome–built environment interactions. The chapter considers a number of factors important to characterizing buildings and microbial communities. It also identifies areas in which progress will be necessary to realize the promise of future studies, including obtaining quantitative microbial information, understanding bioinformatics approaches and assumptions, improving study cross-comparison, and developing datasharing infrastructures. Studies also will be needed that deepen knowledge of how indoor microbial exposures connect to effects on human health, and the chapter examines several approaches to obtaining health-focused information.
A number of different strategies will be needed to incorporate results of such research into integrated built environment–microbiome–human systems and to use the improved understanding of these systems gained from this research to manage indoor built environments to benefit the well-being of occupants. These strategies will necessarily include changing building design and operation, as well as communicating new guidance to occupants, building designers and owners, facilities managers, and others who design, build, occupy, and manage the built environment. The chapter describes those strategies and some of the challenges entailed in their implementation.
A number of studies characterizing microbial communities in different types of built environments have been conducted over the past decade to provide new information about indoor microbiomes. Of interest are not only active microorganisms but also those that are viable but dormant (which may represent a majority of intact microorganisms in a dry building), as well as microbial components and metabolites that may have health effects. Although results are difficult to generalize, these studies provide insight into the relative contributions of various sources to microbial communities in the built environment and similarities and differences among these communities associated with built environment features, along with information on spatial and temporal variation. Prior chapters present a number of these findings, which also have recently been reviewed by Adams and colleagues (2015, 2016) and Stephens (2016). The selection of tools and techniques with which to collect and analyze samples of built environment microbiomes needs to take account of the complexity and variation reflected in these dimensions.
Sources That Populate Indoor Microbial Communities
As discussed in Chapter 3, the microbial communities within built environments derive from a mix of outdoor microorganisms brought indoors through air, water, and occupants and microorganisms from indoor sources. Indoor microbial communities also include microorganisms shed from those who occupy the building, and they are influenced by microbial growth and death driven by indoor environmental conditions, such as the availability of moisture. How these sources contribute to shaping indoor microbial communities varies among buildings and occupants. Because outside air can be an important contributor to indoor fungi, buildings operated with natural ventilation—such as open windows, for example—may have indoor fungal microbial communities that are more similar to those of the environment outside relative to buildings that use mechanical ventilation and air filtration, which can remove some of these microorganisms. Humans reportedly account for 5 to 40 percent of identified microbial sequences across a sampling of built environment studies (Adams et al., 2016), reflecting their role as important but variable contributors to the indoor microbiome. Different materials used within a building also provide surface substrates of differing chemical and physical composition on which microbes settle, although the influence of various surface materials on indoor microbiomes remains unclear.
Spatial and Temporal Resolution
The microbial species and their abundances in outdoor air vary geographically, and they vary frequently by season and potentially over the course of a day. This variation in turn affects the composition of microorganisms entering a building through, for example, ventilation systems. Different areas in a building also show differences in microbial composition, which can reflect various locations within and uses of a room (e.g., floors versus door trim), differences among types of rooms (e.g., a bathroom versus an office), and the likely presence or absence of liquid water (e.g., sinks or showers versus walls of rooms without sources of water). The combination of properties of flooring material and human activity also strongly affects rates of microbial resuspension, which can have a significant impact on the composition of airborne microbes since “resuspended dust is estimated to constitute up to 60% of the total particulate matter in indoor air,” as reported in several studies (Prussin and Marr, 2015, p. 6).
A number of characteristics of buildings and occupants need to be addressed in studies of built environments and their microbiomes as a crucial complement to microbial information. The data collection effort undertaken for the Hospital Microbiome Project illustrates how such measurements can be implemented. This project explored the composition of microbial communities in 10 patient rooms and two nursing stations on two floors of a newly constructed hospital before and after occupancy by patients (Lax et al., 2017; Ramos et al., 2015; Shogan et al., 2013; Smith et al., 2013). As reported by Ramos and colleagues (2015), measurements were taken of more than 80 variables at 5-minute intervals over almost 12 months, including
- indoor environmental conditions, including air dry-bulb temperature, relative humidity, humidity ratio (a measure of absolute humidity or the moisture content of air), and illuminance (a measure of incident light) in the 10 patient rooms and two nursing stations;
- differential pressure between the 10 patient rooms and the hallways;
- surrogate measures of human occupancy and activity in the 10 patient rooms using both indoor carbon dioxide (CO2) concentrations and infrared (IR) beam–break counters installed at the patient room doorways; and
- outdoor air intake fractions in the heating, ventilation, and air conditioning (HVAC) systems serving the two floors.
Accomplishing this data collection required placing a number of commercially available sensors and data storage systems in rooms, hallways, and nursing stations, as well as within HVAC and air-handling units. It also required personnel efforts over the year to check and maintain the sensors and to download and analyze the large amounts of data obtained. Efforts have recently been undertaken to develop an open-source platform that can aid researchers in creating their own systems of linked sensors (such as sensors for temperature, light intensity, and humidity) and data loggers to enable those conducting research on the built environment to design and undertake similar data collection efforts (Ali et al., 2016).
Decisions on which building characteristics to measure and how those decisions are driven by such factors as those highlighted in the prior section reflect the need to capture sufficient information to elucidate microbial sources and to understand factors that support microbial growth and activity, as well as the need to capture sufficient information to account for spatial and temporal variability in buildings and microbiomes. See Box 4-1 for a discussion of the use of longitudinal study designs to understand microbiome–built environment–human interactions.
Within the past 5 years, efforts have been made by those conducting research on the microbiome–built environment relationship to generate a “minimum information standard” for built environment samples. However, questions remain regarding how to balance the collection of sufficient data at sufficient frequency to understand and compare the results of built environment–microbiome studies with the challenges of time, cost, and feasibility. Decisions about which types of samples to collect and which building characteristics to measure also will need to consider applicability to understanding particular health effects. For example, collection of air, water, and dust samples and data on associated building parameters may be important for understanding inhalation exposures.
Building scientists employ a number of techniques to characterize indoor environments. Ramos and Stephens (2014) review many of these techniques, including those for collecting information on “(1) building characteristics and indoor environmental conditions, (2) HVAC system characterizations and ventilation rate measurements, (3) human occupancy measurements, (4) surface characterizations, and (5) air-sampling and aerosol dynamics” (p. 247). Collaborative consortia—such as SinBerBEST, which involves the University of California, Berkeley; Nanyang Technological University; and the National University of Singapore—are also working to develop and improve building sensors to support efficiency, sustainability, and indoor air quality.1 The three areas highlighted below represent
several critical areas and measurement challenges for studies of the built environment.
Measuring Ventilation Type and Airflow Rate
As emphasized in Chapter 3, buildings vary greatly in design and operation, including in the HVAC systems they employ and how these systems are operated, the extent and timing of outdoor air intake versus indoor air recirculation, the ventilation rates, the use and efficiency of air filtration, and other measures. In the context of built environment–microbiome studies, there is no typical building, and one cannot assume ventilation and airflow parameters based only on building type. As a result, information on building systems and ventilation rates is among the data necessary to interpret sample results. In addition to basic characteristics of building
HVAC systems obtained from system specifications, such parameters as airflow rates can be measured in air-handling units, in ductwork, and at room inflow or outflow grills. Tracer gas methods can also be used to measure air change rates and to quantify air distribution, although there are some limitations. CO2 concentrations are commonly measured, for example, but cannot be solely relied on for ventilation estimates (Mudarri, 1997; Persily, 1997).
Measuring Moisture in the Built Environment
As discussed in detail in Chapter 3, the availability of water to microorganisms is a critical factor supporting their growth and activity in the environment. Without available moisture, or at moisture levels too low to support such growth, the built environment can function more like a
“microbial wasteland” (Chase et al., 2016; Gibbons, 2016). Where liquid water is present—for example, in plumbing and in areas that receive continual or periodic wetting, such as sinks—water samples can be collected for microbial analysis, and such characteristics as water temperature, pH, and chemistry can be analyzed. In many sampling locations, however, actual liquid water may not be present.
As noted in Chapter 3, relative humidity is commonly measured in studies of indoor air quality using off-the-shelf commercial sensors. However, relative humidity can vary spatially and temporally in a room and is also not the moisture parameter most relevant to building microorganisms on surfaces. Microorganisms entering through air systems, shed from occupants, or resuspended following human activities gradually settle and deposit onto surfaces.2 A relative humidity measurement taken from the center of a room will not reflect the moisture available to support microbial growth on a surface across the room. The moisture actually present at these surfaces and available to the microorganisms is most relevant to understanding subsequent microbial activity. Moisture content within materials, such as drywall, can be measured using, for example, electrical conductivity sensors, but this internal moisture can be trapped in the bulk material or otherwise unavailable to a microbe. The measure of moisture associated most clearly with microbial growth is water activity (aw) (Adan and Samson, 2011; Dedesko and Siegel, 2015; Flannigan et al., 2011; Harriman and Lstiburek, 2009; Macher et al., 2013). Because aw is difficult to measure directly, however, the most commonly used approximation is equilibrium relative humidity (ERH). This quantity reflects the equilibration of moisture between the air and the material and can be measured using sensors placed in an enclosed space on top of the material, as described by Dedesko and Siegel (2015). ERH has been used in a number of microbiome–built environment studies (e.g., Chase et al., 2016, and many others). It is limited by the fact that it cannot detect the influence of adjacent areas of the material or their interaction with room air.
Measuring Occupancy and Human Activities
The role of humans and human activities in the built environment is complex and combines with building characteristics to affect microbial communities. The density of occupants in the environment affects not
2 The dynamics of particles settling from air is dependent on such factors as the particles’ aerodynamic diameter, which takes into account the effects of density. Different types of microorganisms (such as viruses, bacteria, and single and multicellular fungi) have different size ranges, although settling behavior is likely to be complicated by the fact that these organisms do not exist as single cells in culture, but may be associated in aerosols with dust, soil, water, and other microbial organisms or fragments (Qian et al., 2012).
only microbial shedding but also such variables as room temperature, humidity, and CO2 content, which in turn may impact the operation of building systems designed to maintain environmental parameters and occupant comfort. Walking over different types of flooring resuspends settled microorganisms, as do such activities as vacuuming (Prussin and Marr, 2015; Qian et al., 2014). It is challenging to measure the multiple ways in which humans use indoor spaces. The Hospital Microbiome Project, for example, used such tools as IR beams in doorways to determine how many times people entered or left a room (Ramos et al., 2015). Other measures can be used to gain information on occupants, including activity questionnaires; manual and video observations; smartphone technologies to track and monitor occupants (Zou et al., 2017); and indirect measures of human density, such as indoor CO2 concentrations (also employed in the Hospital Microbiome Project) (Ramos and Stephens, 2014). Approaches that connect chemical signatures with occupants and their activities and improve visualization to help make sense of the large amounts of interacting data and generate new hypotheses may also provide complementary information. Mass spectrometry data on chemicals from personal care products and microbial signatures collected from swab samples have been mapped to room surfaces and occupants to help visualize molecular distributions (Bouslimani et al., 2015; Dorrestein, 2016). Such approaches may represent another opportunity for integration of biological, chemical, computational, and other disciplines in understanding how humans, buildings, and microbial and chemical environments interact. Social and behavioral science research also provides theories and methods useful for studying individual and group behaviors (NASEM, 2017b); further engagement with these disciplines will be particularly valuable in understanding human interactions in built environments and the factors influencing human behaviors.
Building Simulation Tools
A number of design and analysis simulation tools can be useful in understanding the factors impacting environmental conditions that are potentially conducive or unfavorable to indoor microbial growth. Design tools are those used specifically to support building design, while analysis tools are also used to study building performance issues that are not necessarily part of the design process—for example, trying to understand indoor air quality problems in an existing building or to analyze experimental data. Table 4-1 identifies several existing simulation tools that are particularly relevant to addressing building energy use, airflow, contami-
|CHAMPS||Platform for combined building heat, air, moisture, and pollutant simulation and modeling|
|CONTAM||Multizone airflow and contaminant transport software|
|DesignBuilder||Whole-building energy-use simulation tool with graphical user interface to EnergyPlus; includes HVAC system selection and sizing|
|EnergyPlus||DOE’s whole-building energy simulation engine; includes HVAC system selection and sizing as well as code compliance|
|eQUEST||Whole-building energy performance|
|ESP-r||Whole-building energy simulation program for integrated modeling of building energy performance; used primarily to support researchers undertaking detailed studies|
|IDA-ICE||Multizone simulation of indoor thermal climate and whole-building energy consumption|
|IES||Whole-building energy simulation, multizone and CFD airflow, HVAC system design|
|LoopDA3.0||Sizing of natural ventilation openings|
|OpenStudio||Open-source software development kit for building energy simulation|
|THERM||Two-dimensional heat transfer in building components such as windows, walls, foundations, and roofs, providing local temperature patterns that may relate to condensation and moisture damage|
|TRNSYS||Component-based simulation package capable of whole-building energy simulation and design optimization|
|WUFI||Heat and moisture transport through building assemblies such as walls and roofs; capable of one- and two-dimensional analyses|
NOTE: CFD = computational fluid dynamics; CHAMPS = Combined Heat, Air, Moisture, and Pollutant Simulation; DOE = U.S. Department of Energy; eQUEST = Quick Energy Simulation Tool; HVAC = heating, ventilation, and air conditioning; IDA-ICE = IDA Indoor Climate and Energy; IES = Integrated Environmental Solutions; THERM = Two-Dimensional Building Heat-Transfer Modeling; TRNSYS = Transient System Simulation Tool; WUFI = Wärme und Feuchte Instationär.
nant transport, and moisture conditions.3 An important task for future work will be to couple such simulation models for the built environment with explicit population dynamics models (alluded to in Box 1-2 in Chapter 1) for diverse, complex microbial communities. Nonlinear feedbacks within and among interacting populations can lead to surprising effects of
3 More information on these and other tools is available from the International Building Performance Simulation Association (IBPSA) via its Building Energy Software Tools Directory (http://www.buildingenergysoftwaretools.com [accessed May 11, 2017]).
seemingly straightforward interventions in ecological systems, and being aware of such potential outcomes should be part of the conceptual toolkit for understanding microbiomes in the built environment (Abrams, 2009).
In addition to information about building design and operations, studies aimed at understanding the impacts of microbial communities in built environments rely on tools that characterize these microbiomes and their functional activities. The tools and techniques available for this purpose address three main types of questions.
First, which microorganisms are present in the community, and in what quantities? Answering this question requires tools that can detect types of organisms even when they are present at low abundance (sensitivity) and that can accurately detect a target organism in the presence of other types of organisms and confounding material (specificity). It also requires tools that can both identify diverse groups of target organisms to provide information on community composition and yield quantitative information on absolute abundance in the community.
Second, are these microorganisms active, and if so, what are they doing? Addressing this question requires an understanding of whether the microorganisms are capable of replicating (viability), as well as tools that can elucidate their functional activity (biological activity and functional coverage).
Third, what potential do these microorganisms have to cause health effects (negative or positive)? This question connects exposures to microorganisms in the built environment with occupant health effects. In addition to the information on viability, microbial tools that provide information on such molecules as toxins and allergens, epidemiologic investigations, and animal studies (such as dose-response studies) provide important information on relationships between exposures and outcomes (see Chapter 2).
Capabilities of the Current Molecular Toolkit for Characterizing Microbial Communities in the Built Environment
Studying microorganisms by culturing them has been undertaken for decades, but a large majority of microorganisms cannot be cultured, and culture-based approaches generally are too low-throughput to facilitate detailed community-level analysis. “Omics” approaches, a term referring broadly to approaches that yield collective measurements of sets of biologic molecules, have recently come to the forefront for microbial studies. These approaches include genomics (analysis of DNA), transcriptomics (information on mRNA, which reveals which genes are transcribed to be
expressed by a cell), proteomics (data on the suites of proteins produced by cells), and metabolomics (focused on chemical metabolites). Such techniques frequently are used in combination with each other and with culturing to provide information on the presence of different taxonomic groups of microorganisms in a sample and to characterize microbial functions and activities. The parallel processing and computational/bioinformatics algorithms that underpin omics data analysis provide higher-throughput measurements relative to older generations of techniques.
The discussion below starts with a brief review of sample collection, handling, and analysis. It then focuses on information obtained through molecular characterization tools and highlights areas in which future development would support advances in the field. Appendix A contains a more detailed technical analysis of current and emerging molecular characterization tools and qualitative discussion of their performance in terms of features relevant to understanding microbial communities such as those in built environments. These features of sensitivity, specificity, organism coverage, taxonomic resolution, quantification, viability, functional coverage, and reproducibility are briefly introduced starting on page 160.
Sample Collection, Handling, and Analysis
Several aspects of sampling can affect the results obtained in indoor microbiome studies. For example, many studies aimed at characterizing indoor microbial communities employ analyses of microbial nucleic acids. However, recovering microbial DNA or RNA for quantitative analysis depends not only on the sampling method(s) used but also on how samples are handled, extracted, and processed.
Various methods can be used to collect samples from the indoor environment, including air sampling that pulls room air across dry filters or into liquid media, settling onto agar collection plates; vacuum sampling of settled dust; and surface swabbing. Obtaining representative samples of the air, water, and surfaces within the built environment is idiosyncratic, and different sample types have advantages and disadvantages for addressing different types of questions. To further characterize microbial communities and to explore fomite transmission, surface swabs may be useful. To gain insight into human respiratory exposures, however, it may be necessary to collect samples that more closely represent this exposure route, such as air samples or possibly settled dust. The use of long-term composite samples from indoor sources generally is the most advantageous collection paradigm for microbiome characterization. Such samples provide a time-averaged perspective on microbial composition compared with an instantaneous sample, and they may be necessary to obtain sufficient microbial biomass for analysis (Yooseph et al., 2013). Data from so-called grab samples
should not be casually extrapolated into a perspective on exposure. However, which types of samples are most important to collect for purposes of characterizing relevant exposures of building occupants is not fully known, and further work in this area will be valuable in informing the design of future studies to test health connections.
As is the case with the analytical chain of custody for obtaining and handling gas, liquid, or surface samples for analysis of trace chemicals within the built environment, the “ultraclean” practice of using virgin plastics and glassware, all of which must be certified and rendered DNA-free, is essential for any indoor genetic characterizations. As with practices for chemical analysis, analytic blanks and controls (both positive and negative) must be included with the cohorts of indoor environment samples when genetic observations are the goal.
Sample handling and preservation can have a significant effect on the subsequent analysis of genetic material recovered from environmental samples, regardless of microbial origins (viral, fungal, or bacterial). Changes to microbial results have been reported as a consequence of storage conditions prior to analysis (temperature, use of buffers, or other factors). Lauber and colleagues (2010) found for human and soil microbiome samples that “because of the diversity of the samples, conditions tested and analytic methods used, we still lack a comprehensive understanding of how and whether storage of samples before DNA extraction impacts bacterial community analyses and the magnitude of these potential storage effects” (p. 80). Likewise, McKain and colleagues (2013) report changes to the measured proportions of Bacteroidetes versus Firmicutes as a result of differences in sample storage. Discussion continues in the built environment community on optimal sample handling and storage conditions. The most conservative sample preservation methods emphasize immediate storage in ultracold (<–20ºC) and desiccating conditions until the samples can be processed in controlled ultraclean (DNA/RNA-free) circumstances. While it is widely accepted that dry, cryogenic storage (<–60ºC) can preserve genetic materials for relatively long time periods, the shortest possible holding times are preferred prior to the extraction of genetic material from environmental samples. Similarly, extraction of microbial genomes from samples close to the point and time of sample collection is preferred if at all possible.
Diverse extraction protocols and commercial kits are available for recovering microbial genetic material from environmental samples, and recovery of genetic material is rapidly improving with the private sector’s continued development of such protocols and kits. No standards exist in this arena as yet. However, extraction procedures and methods will affect how well microbial nucleic acid is recovered and thus will affect downstream results. As noted in a recent report,
Spores are hardy, for example, but may require aggressive techniques to break them open and release sufficient amounts of an agent’s DNA. Gram-negative bacteria are more easily lysed, but their genomic material also may be more easily sheared and degraded during extraction. As a result, the specific extraction methods used have the potential to bias the types of organisms that will be . . . most efficiently detected. (NRC and IOM, 2015, p. 102)
DNA/RNA extraction practices are expected to continue to develop for the foreseeable future, particularly with respect to recovery from indoor aerosols. Regardless of the diversity of extraction protocols being used, however, the use of parallel internal standards and controls represents best practice for the extraction of genetic material from any given environmental medium.
Identifying and Quantifying the Microorganisms Present in a Built Environment
Sensitivity, Specificity, Organism Coverage, and Taxonomic Resolution
To identify the microorganisms present in a built environment sample, detection technologies need to be able to identify the presence of a microbial group, even when at low abundance. In addition, they need to be able to identify the absence of a particular target when it is not present in the sample. For example, if DNA is extracted from an air filter that also contains pollen particles, a large portion of the recovered genomic material may yield plant genomes rather than the target genomes of the indoor microbial communities (Be et al., 2015).
Sensitivity is a critical parameter for detection, particularly for environmental samples in which many populations exist in low abundance (Rhee et al., 2004; Wu et al., 2001, 2006). Achievable sensitivity also varies for different types of omics technologies. For example, an array-based approach can have an advantage in detecting less abundant organisms compared with a sequencing-based approach (Zhou et al., 2015). Array approaches have not commonly been applied in the built environment setting, however, where studies generally draw on polymerase chain reaction (PCR) and sequencing (methods reviewed in Hoisington et al. ). Improving sensitivity may be a particularly important challenge for the built environment because biomass collected from indoor samples, especially air samples, often is very low. Sample biomass generally is much lower, for example, than that obtained in other types of microbiome sampling for which tools and analysis platforms have been developed, such as samples taken from the human gut. A practical result is that microbial information from air samples
often represents community integration over space and time, because common sampling methods rely on pumps to pull large volumes of air across a filter or analyze samples of dust that has settled over extended time periods.
Taxonomic resolution, on the other hand, is a measure of the information that can be obtained about each microbe in the sampled community (Hanson et al., 2012). Relevant questions include whether a tool enables the microorganisms to be identified at fine resolution, such as the level of an individual strain and species, or at higher taxonomic levels (coarser resolution), such as the level of genus and family. A number of genes are used as phylogenetic markers to provide information on taxonomic groupings. These genes include the 16S rRNA gene for prokaryotes (bacteria and archaea), the 18S rRNA gene for microbial eukaryotes such as protists, and the internally transcribed spacer (ITS) region for fungi. Functional marker genes (such as the genes nifH, amoA, and nirS) can also be used to provide information on microbes in a sample. Analysis of these phylogenetic and functional marker gene sequences often is based on short, several hundred base pair pieces, which limits the obtainable taxonomic resolutions (Jovel et al., 2016; Uyaguari-Diaz et al., 2016). The result is that many microbial ecology studies identify organisms at taxonomic levels coarser than individual species or strains. Some functional markers may be able to provide a finer level of taxonomic resolution than common phylogenetic markers, however.
All tools have strengths and limitations with regard to balancing sensitivity, specificity, organism coverage, and taxonomic resolution. For example, targeted (amplicon) sequencing relies on amplifying a section of DNA from a known gene, which can then be sequenced to obtain more information. The target amplicon can be chosen because it provides information on taxonomic groupings to help classify organisms present in a sample, or it can be chosen to identify a specific known gene. This form of targeted analysis can provide high sensitivity because it can pick up a gene that occurs in only a few organisms in a sample. Another approach to understanding taxonomic composition is to sequence genomic data broadly in a sample (shotgun metagenomic sequencing). Shotgun metagenomic sequence data can theoretically come from any part of each microbial community member’s genome; thus, the information can facilitate tracking genetically differentiated types of organisms, such as strains of bacteria within a species (Greenblum et al., 2015). Metagenomic sequencing and microarray-based approaches can provide greater coverage of different types of organisms relative to amplicon sequencing of a specific gene,4 but their effectiveness
4 A caveat is that amplicon-based sequencing of common marker genes (such as 16S rRNA for bacteria) can provide even broader coverage for the set of microorganisms with that gene (e.g., in identifying bacterial taxa).
presents additional informatics challenges associated with assembly of the wealth of sequence information and new microarray design. The quality and information content of the reference databases needed to make taxonomic assignments represent another important issue. A number of microbial sequences have been deposited in reference databases that do not yet have clear taxonomic classifications.
Relative and Absolute Quantification5
It is useful to know not only which types of microorganisms are present in the microbial communities sampled from a built environment but also how abundant they are. The genomics information obtainable from most studies provides relative quantification, reflecting the abundance of a type of microorganism relative to the total microorganisms measured in the sample (as a fraction of the total). Information on relative abundance can be useful in some characterization studies. However, having information on absolute abundance is also important to enable knowledge to move toward practical application. This information is needed, for example, as part of exposure assessments to better understand dose-response relationships and connections between exposures and human outcomes. Information on both relative and absolute abundance will also be useful in evaluating interventions in the built environment and how they affect microbial communities (a topic discussed further in Chapter 5).
Obtaining absolute quantification information is challenging, however. All measurements are made in comparison with standards, and defining and developing appropriate, validated standards for measurement of microbial communities remains an issue for microbiome studies. Even were such standards available, it is difficult to provide quantitative information with amplicon sequencing and shotgun sequencing approaches (Nayfach et al., 2016; Zhou et al., 2011). Such methods as quantitative PCR (qPCR) would need to be combined with sequencing data to incorporate quantitative or semiquantitative detail. Array-based analysis may also be of use in obtaining quantitative information (Nayfach and Pollard, 2016; Zhou et al., 2015). Moreover, given inherent variations in experimental protocols and bioinformatics analyses, abundance measurements obtained through omics technologies can differ among samples even under identical conditions. There are also technical challenges associated with environmental sampling
5 Relative abundance of a microorganism in a sample refers to the percentage of that type of microorganism that was identified relative to the total microorganisms identified in that sample. Absolute abundance, on the other hand, would reflect the actual number of that type of microorganism that was in the substrate (surface, air, water, or bulk material) in the built environment from which the sample was collected.
and nucleic acid extraction that make it difficult to ensure that the genomic information obtained matches the communities that exist in the original built environment. A variety of interfering substances, such as chemicals in dust, can be present in built environment samples that complicate the ability to successfully extract and amplify DNA or that have genomic material of their own (such as pollen) that may make it challenging to pick up information from rare microbial species. In addition, it is easier to obtain DNA from some types of microorganisms than others using standard extraction protocols, a factor that affects the abundances of microorganisms detected in the sample (Peccia and Hernandez, 2006). A further biological issue, particularly for fungi, is that microorganisms may vary in the number of rRNA copy numbers present per cell, making it challenging to obtain accurate information if such genes are used to assess population numbers (Taylor et al., 2016).
These issues all represent important impediments to the ability to relate microbiome data to fundamental models of population, community, and ecosystem ecology. They also hinder health risk assessment, where absolute numbers matter.
Understanding the Viability and Functions of Microorganisms
Information about the viability and functional activities of indoor microorganisms can be obtained in several ways. Viable microorganisms are those that maintain the ability to replicate in the built environment under suitable environmental conditions. Because the built environment generally has limited moisture, particularly in the air and on surfaces, some microorganisms may be viable but inactive (e.g., as fungal spores) until conditions change. Alternatively, microorganisms detected by DNA sequencing may be “dead” or may exist only as partial microbial fragments and components. The traditional approach for assessing viability has been to culture microorganisms—for example, on agar culture plates—and this remains the standard for pure culture experiments or research involving known pathogens or specific microorganisms that can be cultured in these ways. Yet, while culture-based approaches remain an important complementary technology to such tools as genomics, suitable culture conditions for many microorganisms are not known, or do not account for the range of environments that prompt microbial metabolic activity and reproduction, or may miss the activity of less abundant taxa. Culture-based measures thus are limited as to the information about microbial communities they can provide.
Crucially, researchers also need to know what the microorganisms in a built environment are doing. The biological activity of a microorganism refers to its metabolic functions as a living entity, regardless of its ability to propagate. As discussed above, metagenomics can provide a “snapshot”
of the diversity of a microbial population and thus some information on functional potential in a community, but this DNA-based information does not reveal whether the microorganisms are actively engaged in metabolic activity. Other omics approaches, including metatranscriptomics, metaproteomics, and metametabolomics, provide additional information for characterizing microbial communities functionally (Gutarowska et al., 2015; Zhou et al., 2015).
Additionally, microorganisms produce many molecules that can be measured in the built environment. These include bacterial endotoxin, fungal mycotoxins, and a variety of bacterial and fungal cell wall components that may have toxic effects on cells as a result of exposure. A variety of microbial volatile organic compounds can also be measured in indoor air (Araki et al., 2009). Such techniques as mass spectrometry and associated variations, used to analyze chemicals based on their mass and charge, are valuable for identifying microbial molecules in buildings (Saraf et al., 1997). Measuring these molecules in the built environment can provide markers for the presence of the respective microorganisms (which may not be culturable), as well as yield information relevant to understanding the potential health effects of bioaerosols and microbial samples collected from the built environment. For example, Bordetella pertussis, the airborne bacterium responsible for whooping cough, has not been recovered from ambient aerosol by conventional culture techniques, but it produces an exotoxin that can be measured (Yao et al., 2009). Where particular molecules can be linked to health or other effects, it may be possible to incorporate future monitoring. The use of microbial molecules as markers also has some limitations, since a given molecule may be produced by multiple microbial species, limiting taxonomic resolution. Ergosterol, for example, has been used as a surrogate measure of fungal biomass. Some molecules also may be carried in or persist in the environment even if the producing microorganism is no longer present.
Ideally, a combination of DNA- and mRNA-based measurements, as well as protein- and metabolite-based measurements, would be used to assess the presence and activities of microbial populations in a community in complementary and mutually reinforcing ways. Understanding the functional activities of microorganisms in microbial communities in a built environment remains a particular challenge. Because the functions associated with particular genes and molecules may remain unknown, even recovery of the complete complement of proteins, genes, or metabolites does not automatically yield an accurate functional assessment. However, integrating molecular data from multiple sources—genome, transcript, protein, and metabolite—presents an important opportunity to identify more accurately the biological processes that explain how the diverse elements of
Reproducibility and Development of Reference Materials
Understanding variability and reproducibility among microbial samples in a built environment is a challenge for reasons noted, including low biomass, environmental conditions, and an imperfect ability to extract and measure information and link it to microbial species. However, understanding the strengths and limitations of existing studies will enable comparisons across study results. Given the high community complexity and potentially dynamic nature of microorganisms, natural biologic variation can prevent two laboratories from producing the same results even when they control for technical variation. For example, microscale variation in the composition of microbial communities may result in differences (Kauserud et al., 2012; Pinto and Raskin, 2012; Zhou et al., 2011) even among samples collected by the same laboratory from proximate locations. In genomics studies, part of the reproducibility challenge is also due to technical variation and to the technologies themselves because of inherent measurement errors and biases. Part of the variation among samples can be laboratory-based as well, resulting from differences in sequencing depths (Bartram et al., 2011; Lemos et al., 2012; Zhou et al., 2011) or from variations in sequencing and sequence preprocessing approaches (Pinto and Raskin, 2012; Schloss et al., 2011).
To help address reproducibility and cross-experiment comparison, several benchmarking efforts and efforts to develop reference microbial communities have been and continue to be undertaken. The National Institute of Standards and Technology (NIST), for example, has been active in efforts to standardize microbiome measures, although attempts to apply this work to the built environment remain nascent (NIST, 2012).6 Nascent efforts are also focused on designing mock microbial communities and benchmarking standards. Opportunities to develop reference materials that better capture living biologic material in a controlled environment will further enhance existing reference material resources. For example, a recent effort—Mock Bacteria ARchaea Community (MBARC-26) (Singer et al., 2016)—involves attempting to construct a mock community with representation from environmental habitats, although this work does not encompass the built environment. The EcoFAB initiative being carried out through Lawrence Berkeley National Laboratory is also focused on
6 See https://www.nist.gov/news-events/events/2016/08/standards-microbiome-measurements-workshop (accessed July 16, 2017); https://www.nist.gov/programs-projects/microbiome-community-measurements (accessed July 16, 2017).
developing model microbial ecosystems to improve understanding of microbial communities in humans and animals and in environments such as soil (Berkeley Lab, 2015). Although this effort has not yet incorporated the built environment, this may be an area for further development. In addition, NIST recently cofounded the International Metagenomics and Microbiome Standards Alliance, which may serve as a consortium to help organize future efforts in designing standards and reference materials. It is important to note that model microbial communities may not capture all in situ interactions and activities, and validation in human populations and built environments will be necessary to confirm their utility. However, efforts to define and develop mock communities for the built environment would be helpful in improving standard metrics and benchmarks.
Open questions remain with respect to how best to build on existing benchmarking efforts to guide validation, modeling, and cross-study comparisons in support of future work on the built environment. Developing a better understanding of the conditions under which accurate and reproducible microbiome measurements in built environments can be made will be a foundational requirement for moving investigative research in this area toward practical applications.
Gaining a holistic picture of how the design and operations of built environments, the identities and functions of microbial communities, and potential impacts on humans and the environment are interconnected will require new studies and study designs, particularly as research moves beyond ecological characterization and toward translation and application.
Elucidating Causal Connections Between Microbial Exposures and Human Health Outcomes
Supporting or promoting health is a key motivation for understanding indoor microbiomes and using that knowledge to inform how buildings are designed, built, maintained, and operated. To make progress toward such practical applications, researchers will need to build on the existing base of studies to develop and test hypotheses. A number of steps are needed to determine the public health relevance of interrelationships among built environments, indoor microbiomes, and humans. The general steps summarized below involve the collection and analysis of data in a manner aimed at demonstrating relationships in a clinically relevant framework. The committee is not suggesting that all of these types of information need to be collected for all studies, but rather that these are considerations:
- Define the objective to be tested in the study, such as the hypothesis that an indoor microbial exposure relates to a certain health outcome.
- Identify the microbial exposure or exposures of interest by taxa (where they are on the biologic tree of microbes) and by function (what are they doing individually and together). For example, are microorganisms producing molecules that could interact with human cells in the airways? It will be important to identify techniques for measuring these microbial exposures and their functions.
- Identify clinically relevant measures of the health outcomes being assessed.
- Identify features of the built environment hypothesized to be relevant through their effects on indoor microbial communities or their direct impacts on human health. Identify strategies and techniques for measuring these features of the built environment.
- Identify the relationships among microbial exposures, features of the built environment, and relevant health outcomes. These relationships may vary for different types of microorganisms or microbial communities, types of exposure, doses and stages of life, individual susceptibility, such cofactors as stress, or other factors.
- Collect information on appropriate building data, such as temperature, light, airflow, and other building system characteristics. Also collect information on such occupant factors as activities, cleaning practices, health status, and social factors.
- Analyze the data in a manner that sheds light on the plausibility, consistency, and reproducibility of results. For example, longitudinal data may highlight an increased risk associated with higher baseline exposure. If feasible to study, does the outcome change if the exposure being tested is removed? For certain types of exposures, it may be possible to establish dose-response relationships through appropriate study design.
- Examine potential confounders and effect modifiers to understand their role in observed associations between built environment microbial exposures and human health outcomes. Continue the process until the totality of evidence is strong enough to support informed decision making.
Uses and Limitations of Epidemiologic and Animal Studies in Understanding Health Effects
A variety of study designs can be useful in increasing evidence for correlations already suggested between built environment microbial exposures and health outcomes, in generating evidence to support or refute additional links, and in extending observed associations and animal model results to
causal connections in humans. Generating evidence may require iterations between longitudinal epidemiologic models in humans, validation in animal models, and testing through intervention studies, described briefly below:
- Longitudinal cohort (observational) studies: These studies follow groups of subjects, ideally comparing those exposed and not exposed to a hypothesized risk factor with respect to the occurrence of a health outcome. A subtype of studies follows individuals from around birth (birth cohort studies), which can be particularly useful in understanding effects of early-life exposures on health later in life. Longitudinal cohort studies are preferred over retrospective case-control or cross-sectional studies because they can be used to identify correlations between exposure and disease over time. Investigators may take advantage of longitudinal birth cohort studies to assess, for example, whether studies showing protective properties arising from exposure to microbes in farm environments are reproducible or generalizable to other settings. The new U.S. national consortium on birth cohort studies (Environmental Influences on Child Health Outcomes [ECHO]) may offer one type of study infrastructure that can provide opportunities to facilitate such research.
- Animal validation studies: Controlled studies in animals can be useful in testing and refining observational correlations. One study, for example, tested the observation that exposure to a diverse microbial community was associated with reduced allergic responses by feeding dust from a house with a dog to mice and then studying changes in their gut microbiomes and how those changes affected immune response (Fujimura et al., 2014). Animal studies can both provide greater control over exposures and environmental conditions relative to human observational studies and help establish important dose-response relationships.
- Intervention studies: Such studies categorize participants into groups that do or do not experience a particular intervention to examine its effects. For example, one can envision built environment interventions that change cleaning practices, water temperatures, indoor humidity levels, or other building systems and conditions and occupant behaviors in order to test whether these changes affect a heath outcome of interest. Given the multitude of factors influencing microbial communities in buildings and human exposures to those communities and the difficulties of trying to disentangle effects on health, intervention studies can be useful in further testing relationships hypothesized or observed through observational studies and in animal models.
Testing a well-defined exposure and well-defined health outcome for which clear assessment measures are available improves the statistical power of a study. However, epidemiologic and animal studies have important limitations when applied to the built environment field. In this field, data collection will incorporate not only the building and microbial features discussed in earlier sections (such as ventilation systems, chemical emissions from building materials, physicochemical variation in the built environment, and microbial proliferation), it also will take into account occupant factors that can complicate results. Applying epidemiologic and animal studies to understand exposure to a specific microorganism, such as a pathogen, is more straightforward than applying such studies to tease out multiple microbial exposures, such as those that occur in most built environments. For example, to understand how characteristics of the built environment and its microbiome influence childhood neurodevelopment and neurocognitive outcomes, it will be essential to consider the sample types that should be collected and the microbial parameters that should be measured to characterize the relevant exposures. It will also be essential to consider the roles of factors beyond microbial exposure, including socioeconomic status, diet, environmental pollutants, and such social parameters as educational background. For noninfectious outcomes, there may be substantial time between an exposure and its observed effect (e.g., if there is an important early-life window that influences later development). And people generally do not experience a single built environment but rather multiple environments as they move from home to car to office to gym or movie theater and back to home. Deriving associations among this array of factors will be highly challenging, as will identifying their relationships to specific outcomes.
Both strategies for improving the integration of such data into estimates of disease burden and the use of appropriate sensors, which can serve as automated endpoint technologies for monitoring of building microenvironments and other factors, can contribute to meeting these challenges. A number of sensor technologies already exist, although greater use of personal sensor systems could be explored to enable improved monitoring of the personal activities and exposures of individuals under study. Observational studies with longitudinal and cross-sectional analyses can help define physical, chemical, and biological markers with associations with health outcomes. Identifying the microbial and metabolic biomarkers associated with disease burden, disease onset, and disease or treatment outcomes will be particularly important in connecting environmental microbiome exposures to health effects. These biomarkers and their mechanisms of action can be one focus of future studies under controlled conditions (such as animal studies) to gain clearer understanding of specific microbial exposures, dose-response relationships, and physiologic outcomes. Further improve-
ments in such areas as transport modeling can also contribute. Together, these elements can allow for studying components of systems biology when multiple interacting effectors and outcomes are involved. Teasing out these relationships may require network and machine learning approaches (and additional statistical tools), which would help support further development of the field, although important work remains to be done in validating such approaches for built environment analyses.
Assessing the Utility of Prior Study Data and Stored Samples
Some samples currently in storage from prior studies conducted for other purposes might be useful resources for the microbiome–built environment research community. For example, the 2006 National Health and Nutrition Examination Survey (NHANES) of the National Children’s Study (NCS) archive included dust and blood samples along with questionnaires covering a variety of health issues.7 To understand whether and how prior study samples could be reanalyzed, the community will need to define the characteristics that determine sample quality and utility with regard to the particular study questions being asked. For microbiome analysis, a sample needs to have been handled and stored appropriately, and its utility can be assessed qualitatively based on the expectation for microbial profiles associated with similar built environments. Having a high-quality sample may not be absolutely necessary if the degradation of the community signature was not so complete that it impaired the ability to detect trend differences among different conditions. Care must be taken, however, to ensure that the sample is sufficient to address the hypotheses being tested and the relevance of the findings can be confirmed. In addition, the anticipated associations and evaluation of sample-derived data are based on current assumptions, which may not turn out to be appropriate. Sample utility is also affected by knowing whether the associative variables, such as health outcomes and building measurements, were collected appropriately to enable testing of associations with the microbial and metabolic profiles that can be derived from these samples. For example, asthma research studies often are well suited to helping to derive health outcomes associated with microbial profiles, but studies on dust chemistry and childhood development are likely to have limited applicability because of a lack of control associated with variable measurement and a sparse matrix of sample acquisition that reduces the relevance for interpretation of microbial exposure.
7 Information on the NCS repository is available at https://www.nichd.nih.gov/research/NCS/Pages/researchers.aspx#data (accessed July 16, 2017). Information on dust samples collected as part of NHANES is available at https://wwwn.cdc.gov/nchs/nhanes/2005-2006/ALDUST_D.htm (accessed May 11, 2017).
One important aspect of assessing prior studies for potential analysis of microbiome–built environment samples is determining the list of collected variables that could be potential confounding influences for the particular outcome of interest (e.g., pet ownership, household occupancy, diet and lifestyle, educational profiles, pollution exposure), and these relevant factors may not have been recorded in prior studies. As a result, applying new microbiome analysis to previously collected samples may be possible but will likely be restricted to testing hypotheses contextualized by the original focus of the study, particularly for case-control studies. Thus, population-based epidemiologic studies that have collected a large amount of exposure and outcome information, which can be analyzed in combination with analysis of microbial samples, are likely to be most useful.
Taking Advantage of Near-Term Opportunities
While time, effort, and significant fiscal commitment from public and private entities will be required for many of the areas of investigation identified in this chapter to come to fruition, the community of microbiome–built environment researchers could leverage near-term opportunities to study linkages among building conditions, building microbiomes, and humans. For example, opportunities for such studies may arise during disaster response and in efforts to support resilience planning. Acute events such as Hurricanes Katrina and Sandy resulted in widespread flooding of homes and underground subway stations, and they provided an opportunity to examine how microbial communities changed when exposed to these extreme conditions. The U.S. National Response Team,8 which “provides technical assistance, resources and coordination on preparedness, planning, response and recovery activities for emergencies,” and NIST investigative teams that enforce the National Construction Safety Team Act, which “establishes investigative teams to assess building performance and emergency response and evacuation procedures in the wake of any building failure,”9 could provide potential opportunities for the involvement of researchers exploring microbiome–built environment–health interactions by fostering appropriate agency and organizational connections.
9 Public Law 107-231, October 1, 2002 (https://www.gpo.gov/fdsys/pkg/PLAW-107publ231/pdf/PLAW-107publ231.pdf [accessed May 11, 2017]).
Obtaining Quantitative Microbial Information
As discussed above, many biologic and technical challenges are entailed in obtaining data on relative and absolute abundance for built environment microbiome samples. Achieving advances in methods for improving the quantitative information that can be obtained from samples is important for future progress in the field. Not only is quantitative information important for establishing connections with health outcomes, it also helps underpin the development and interpretation of models for potential interventions in the built environment and analysis of their impacts.
Improving Comparison Across Studies
Multiple challenges are also entailed in drawing more effective comparisons across the results of existing microbiome–built environment studies. Different groups may collect samples from the built environment in different ways, may collect different sets of building and occupant data to accompany samples, may use different characterization tools, or may undertake sample analysis and data interpretation differently. Genomics tools, for example, identify a diverse set of bacterial organisms in the built environment, and the number of catalogued organisms continues to grow as databases are updated to reflect ongoing microbial sequencing. Several different databases are used in this context, each of which has different levels of deposit criteria, quality control, and curation. However, the reproducibility of measurements of microbiome composition can also vary depending on experimental conditions. The ability to compare results across studies enables researchers to better assess converging lines of independent evidence in parsing the factors that affect the formation and functions of microbial communities in built environments. Efforts to generate standards for the collection of data and metadata, common storage formats and resources, and microbial reference materials that can be used to calibrate results across laboratories are important in addressing this need. Galaxy (Afgan et al., 2016) and Cyverse (Merchant et al., 2016) are examples of ongoing efforts to build software and infrastructure for complex and computationally expensive omic analytic pipelines that are intended to be scalable, shareable, and reproducible through the use of version control, detailed scientific workflow10 logs, common data standards, and access to large-scale computing resources. These
10 Additional background on scientific workflows can be found at http://cnx.org/contents/j-3C75Ok@3/Scientific-Workflows (accessed July 16, 2017).
examples demonstrate the work needed to generate reproducible analysis with microbiome–built environment datasets.
Understanding Bioinformatics Assumptions and Limitations
The omics technologies now used to help characterize microbiomes rely on bioinformatics algorithms to make sense of the information obtained and to determine the reference databases needed to link, for example, the obtained sequence information to the taxonomic identification of the microorganisms detected in built environment samples. Sequence data generally are interpreted using programs that compare sequences for similarity and assign them to operational taxonomic units (OTUs) based on a predesignated similarity for a given level of ecological resolution (such as 97 percent similar). Different underlying assumptions encoded in the informatics software, however, can affect how these sequences are clustered into OTUs and how they result in particular microbial taxonomic assignments. Efforts to develop simulated datasets with which to benchmark computational tools have demonstrated that different descriptions of community structure may be found for the same input data depending on the analysis tool and parameters selected (Lindgreen et al., 2016; Randle-Boggis et al., 2016; Weiss et al., 2016).11 Understanding the assumptions underlying bioinformatics software and how results compare across different informatics packages is also foundational in understanding built environment–microbiome results and moving toward their application. The further development and adoption of benchmarking and reference standards will be valuable in this regard. This point applies to the methods used to recover DNA from environmental samples, as well as the primers used to amplify specific DNA sequences, which typically are used as ecological identifiers for phylogenetic comparisons. Also important is requiring that the software tools and detailed scientific workflows used to generate an analysis be made available for peer review to help ensure that results are independently verifiable.
Supporting Sharing of Data on Microbial Communities and Metadata on Buildings and Building Systems Through the Use of Data Commons
A data commons is a collection of computational resources that provide a common platform for access to data sources for analysis, supporting a community of researchers. Key components include storage for generated data, metadata integration, and retrieval of data in forms that enable downstream analysis. Developing this resource for microbiome–built environ-
ment research will require publicly accessible data repositories, such as the Sequence Read Archive, where raw genomic data or raw metabolomics data (Wang et al., 2016) can be housed. It also will require data-sharing standards to ensure sufficiently complete and accurate descriptions of the experimental conditions used to generate new microbiome data (Leinonen et al., 2011). Ensuring that experimental data and software are accessible to the research community is key to supporting the microbiome–built environment field.
A data commons can be used to ensure that differences in results obtained with computational analysis tools are understood more clearly by enabling comparisons across common datasets. Data commons also encourage the development of new modeling or analysis tools by providing data access to tool developers who are not data generators using common data formats. Moreover, microbiome–built environment data are collected from a diverse array of research efforts across multiple institutions and research disciplines, all with the need for analysis tools that can operate on these data collections using a potentially distributed set of computing resources to support scalable and independent analysis, and in ways that enable individual investigators to contribute to the field as a whole. A data commons can also provide access to common analytic pipelines, the contents, logic, and algorithms of which are public and that provide standards for analysis, such as in community profiling.12
A previous effort to develop minimum requirements for building metadata led to a defined set of data to be collected in conjunction with experimental studies of indoor microbial communities—the MIxS-BE package. This standard currently includes parameters such as air and surface temperatures; measures of air and surface humidity; surface type, material, and pH; type of HVAC and filter system; and a number of other details (Glass et al., 2014).13 While commendable in addressing the need for such information, these data are also somewhat limited in detail and extent. For example, the metadata template includes type of heating and cooling system, but the listed options encompass no information on how the system is configured or controlled or other key information needed to characterize these systems more fully. Similarly, indoor surfaces are described in terms of location or type but without greater detail on the material or its likelihood of retaining moisture.
12 Examples of data commons that help support various research communities include National Institutes of Health (NIH) metabolomics and microbiome data-sharing requirements (http://www.metabolomicsworkbench.org/nihmetabolomics/datasharing.html [accessed July 25, 2017]); the Nephele cloud-based microbiome analysis pipeline (https://nephele.niaid.nih.gov [accessed July 25, 2017]); the Stanford Data Science Initiative (https://sdsi.stanford.edu/data-commons [accessed July 25, 2017]); MassIVE (http://gnps.ucsd.edu [accessed July 25, 2017]); and QIITA (http://qiita.microbio.me [accessed July 25, 2017]).
A similar building and system data definition effort was undertaken in conjunction with the U.S. Environmental Protection Agency (EPA) Building Assessment Survey and Evaluation (BASE) study of indoor air quality conditions and outcomes in 100 U.S. office buildings (EPA, 2006). That study included protocols for collecting more detailed information than called for in MIxS-BE on the building, the spaces being studied, and the HVAC systems serving those spaces, as well as for conducting measurements of environmental conditions and ventilation system performance. It addressed the condition of many system components, including their functions, state of repair, and dirt and moisture levels, and it contained a more detailed description of HVAC system type and control. The BASE protocol is more informative than the MIxS-BE package, although it may be more detailed than is needed for many indoor microbial studies. Given that the BASE protocol was developed more than 20 years ago, it may be useful to update it for its potential application to indoor microbial studies and to continue efforts to develop common data templates. The trade-offs between obtaining as much useful information as possible to characterize buildings, occupants, and their environments and the volume of information to be collected and analyzed also will require further discussion and agreement. An updated and consensus-oriented protocol for building and system characterization would be useful for future studies of the design, condition, and performance of buildings to advance understanding of a range of indoor microbial issues.
Thus, opportunities to refine data specification frameworks are needed, as are efforts to ensure that data can be accessible across publicly searchable databases, which have open curation standards. For sequence information, databases include those maintained by the European Bioinformatics Institute (EBI) and the National Institutes of Health’s (NIH’s) U.S. National Center for Biotechnology Information (NCBI), although efforts to make nongenomic metadata widely available through central repositories appear to be less well developed (Dorrestein, 2017; Vizcaíno et al., 2016; Wang et al., 2016). Long-term support is needed for large-scale data repositories that store both raw and processed forms of omics data. Storing raw data output will be critical to ensure that more complete and accurate recovery of genomic, proteomic, and metabolomic data can be obtained through algorithmic improvements well after the data have been collected. All of these efforts will require the engagement of researchers with built environment, building science, and engineering expertise, along with microbial ecologists and other researchers that are experts in microbiome measurement data (Abarenkov et al., 2016). Potential partners for further development of a data commons may exist in multiple federal agencies, including NIST, EPA, the National Institute for Occupational Safety and Health (NIOSH), and NCBI; professional societies, including the American Conference of Governmental Industrial Hygienists (ACGIH), the American Society
of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE), the American Society for Microbiology (ASM), the Indoor Air Quality Association (IAQA), and the International Society of Indoor Air Quality and Climate (ISIAQ); and others.
The results of microbiome–built environment–health research can be translated into practical action through multiple avenues, reviewed below.
Effecting Change in Building Practices
As new levels of understanding are achieved, a variety of strategies can be used to translate that knowledge into practice in building design and operation. These strategies are listed from fastest and easiest to those that will be slower to implement but may have the broadest impact.
- Best practices for building design and operation can be described in reports and other documents written by researchers and other experts and intended for practitioners.
- Voluntary guidance on how to design and operate buildings to support improved indoor microbial environments can be produced by engineering and professional societies, such as the American Industrial Hygiene Association (AIHA); ASHRAE; the Building Owners and Managers Association (BOMA); and government agencies such as EPA, the General Services Administration (GSA), NIOSH, and state public health departments.
- Voluntary building rating, labeling, and certification programs and green building programs such as Leadership in Energy and Environmental Design (LEED), Green Globes, WELL, and Fitwel may be particularly appropriate in the context of attempting to advance practice toward a higher level of building performance.
- Industry consensus standards for the design and operation of buildings to support improved indoor microbial environments, written by standards development organizations such as ASHRAE, include minimum standards intended for wide application. They include ASHRAE Standard 62.1 (ventilation for indoor air quality) and standards specifically directed at high-performance buildings such as ASHRAE/ Illuminating Engineering Society (IES)/U.S. Green Building Council (USGBC) Standard 189.1 (on design of high-performance green buildings).
- Model building codes, such as those promulgated by the International Association of Plumbing and Mechanical Officials (IAPMO),
the International Code Council (ICC), and others, generally are based on consensus standards and subsequently adopted by local jurisdictions in their building codes and regulations (often with modifications based on local needs and priorities). Similarly, requirements for building design and operation can be developed for federal agencies that design and operate their own buildings, such as the U.S. Department of Defense (DOD) and GSA.
Standards and Their Limitations: The Example of Ventilation
One of the approaches listed above that can help effect change is embedding knowledge in standards. The section reviews the uses and limitations of ventilation standards as an example.
Ventilation requirements in standards and building regulations are essential to the design of buildings, yet they have several limitations. For example, using the same outdoor air requirement for all spaces of the same type ignores important differences among occupants and their activities, the materials and furnishings in the spaces, and the quality of the outdoor air. These requirements are also intentionally minimum values, meaning that anything lower would violate the standard or regulation. In practice, these minimum values are used without considering the potential benefits of higher levels of outdoor air intake.
It is important to bear in mind that while ventilation requirements are important for establishing building design goals, they are only a first step in the process of achieving effective ventilation in buildings. Once the outdoor air (and exhaust) ventilation requirements for a building and its spaces have been determined, these requirements need to be incorporated into the design of the building and its ventilation systems and properly documented so contractors and building operators will understand the assumptions on which the design is based. Following design, the system needs to be properly installed and commissioned to ensure that it is complying with the design requirements. This latter step involves testing the installed system under a range of operating conditions, including different internal loads, control sequences, and outdoor weather conditions. Finally, the system needs to be operated and maintained effectively over the life of the building to ensure that the system continues to perform as intended. System operation and maintenance involves periodic inspection of system components, calibration of sensors used to control the system, and many other steps, which are described in standards and other documents (ASHRAE, 2012, 2016). Given that building and space uses (including occupancy) often change during the life of a building, it is critical that the ventilation design requirements be reevaluated when such changes occur to ensure that the system can continue to meet the building’s ventilation needs. These points highlight the fact that
even after knowledge needed to support changes in practice has been identified and translated to formal requirements, many other factors involving building designers, owners, operations and maintenance personnel, and occupants come into play.
Communication and Engagement
Designers, professional societies, owners, operators, and occupants all need to be engaged to support an effective translation of built environment–microbiome information from research into practice. The results of studies on the microbial communities that surround people every day in their homes, schools, and offices and what impacts these communities may be having on people’s health and well-being can be of wide public interest. It will be important to communicate effectively about the results of ongoing studies, as well as the caveats on and limitations of that knowledge. Communicating to people that they are surrounded by microbial communities whose effects may include beneficial, neutral, or harmful interactions and providing people with information they need to make choices about their built environments are important goals. At the same time, investigators will want to avoid promoting unjustified fears about the microbial ecosystems that coexist with humans or overselling the strength of available evidence.
The importance of public engagement with and communication about science continues to gain recognition. A recent report from the National Academies of Sciences, Engineering, and Medicine notes that those communicating about science need to consider the goals of the communication—for example, whether it is intended primarily to provide information or to influence behavior—and to align the communication approaches used accordingly. The report also emphasizes the limitations of the “deficit model” of science communication and makes suggestions for a research agenda to improve effective science communication practices (NASEM, 2017a). The “deficit model” assumes that if people only had more factual information about a topic, they would behave in a manner consistent with the scientific evidence. This model has repeatedly been shown to be wrong; however, decision making and behavior are influenced by many factors other than scientific evidence. Translating the findings of microbiome–built environment research into policy and practice will require not only the integration of scientific and clinical information but also consideration of such factors as economic costs and benefits, personal values, and social and political realities. Different potential audiences may be interested in what they can do based on the knowledge communicated, but they are likely to have varying resources, values, and competing priorities (Kahlor, 2016). The involvement of experts from the social and behavioral sciences in microbiome–built environment–health studies can be a useful
strategy for elucidating the many factors relevant to stakeholder communities and effectively designing and undertaking engagement.
Citizen science efforts are another approach to building interest in and awareness about scientific topics, and the microbiome–built environment field is well suited to such efforts. For example, the Wild Life in Our Homes project describes the hypotheses being tested in understandable language, invites people to collect and send samples from their homes, and has analyzed the microbial diversity those samples contain.14 Similarly, Project MERCCURI (Microbial Ecology Research Combining Citizen and University Researchers on ISS) sent publicly collected microbial samples into space and examined their growth (Coil et al., 2016).15
Useful suites of tools exist with which to characterize building and occupant factors and microbial communities. Researchers can draw on a variety of omics tools and bioinformatics approaches to characterize indoor microorganisms and study their activities. In addition, various sensors and simulation tools are available for gathering data on building systems and occupant activity.
Both experiments and modeling will help the research community better understand the interrelationships among buildings, microbial communities, and human occupants, and this understanding will support eventual application of the knowledge gained through research. Information from complementary approaches identifying and characterizing indoor microorganisms and microbial products, describing building parameters, capturing occupant behaviors, and collecting human exposure and health data will need to be integrated by the field.
Further efforts in foundational areas that support the research infrastructure for built environment–microbiome studies are needed. The research infrastructure that supports the field includes components that affect the ability of investigators to collect, analyze, store, share, and compare information. Important aspects of this system include continued improvements in microbial and building characterization tools, data collection standards, reference materials, and benchmarking efforts, such as the development of mock microbial communities, validation of experimental approaches, and resources for accessible data storage and sharing to facilitate cross-study comparison and the generation of new hypotheses.
14 See http://robdunnlab.com/projects/wild-life-of-our-homes (accessed May 11, 2017).
Interest in connecting microbial characterization of the built environment to an improved understanding of human health impacts will benefit from studies designed to address health-relevant hypotheses. A number of considerations need to be incorporated into the design and conduct of studies aimed at clarifying potential health effects. These considerations include identifying and collecting the types of built environment, microbial, and occupant samples and data most relevant for understanding exposures; identifying appropriate measures for assessing the health outcome(s) of interest; and developing improved and validated approaches to exposure assessment. A variety of study types, including observational epidemiologic (longitudinal) studies, animal model studies, and intervention studies, will be useful.
Many groups are involved in conducting microbiome–built environment research and moving the knowledge thereby gained toward practical changes in such areas as building and indoor air quality codes and standards. The communities that will need to be engaged in this process include building, microbial, and clinical and public health scientists conducting investigations; chemical and materials scientists; building designers; and communities of professional practitioners. Once sufficient knowledge has been gained, a number of strategies can aid in translating that knowledge into practice, from voluntary guidance and descriptions of best practices to formal codes. However, developing the actual and virtual infrastructures needed to promote effective interdisciplinary research and communication in the field will require sustained engagement and funding.
On the basis of the summary observations above and the information developed in this chapter on the sets of available tools, the committee identified the following goals to address capability gaps and advance the field:
- Develop the research infrastructure in the microbiome–built environment–human field needed to promote reproducibility and enhance cross-study comparison. A framework for establishing further infrastructure to support this field will usefully include the development of shared understandings among investigators on sample and metadata collection and on sample handling, storage, and processing conditions to support effectively addressing different types of research questions, along with the promulgation of best practices and metrics for analysis. The research infrastructure will need to encompass the use of a variety of complementary experimental, modeling, and analysis tools to understand the composition and function of microbial communities and to connect such research to
impacts of interest, such as human health effects, materials degradation, changes in energy usage, and others. The development of community standards, reference materials, and benchmarking materials will be valuable, as will the development of a broader data commons that includes the ability to share data in accessible ways to facilitate integrated data analyses and cross-study comparisons. These efforts are not trivial, and the committee does not mean to imply that the community is unaware of these needs, but to highlight that further progress in establishing this fundamental infrastructure will contribute to the advancement of the field.
- Develop infrastructures and practices to support effective communication and engagement with those who own, operate, occupy, and manage built environments. This will be an important area for attention, especially as the field continues to advance toward application. Social and behavioral scientists expert in such areas as communication can provide insights to inform these efforts.
Abarenkov, K., R. I. Adams, I. Laszlo, A. Agan, E. Ambrosio, A. Antonelli, M. Bahram, J. Bengtsson-Palme, G. Bok, P. Cangren, V. Coimbra, C. Coleine, C. Gustafsson, J. He, T. Hofmann, E. Kristiansson, E. Larsson, T. Larsson, Y. Liu, S. Martinsson, W. Meyer, M. Panova, N. Pombubpa, C. Ritter, M. Ryberg, S. Svantesson, R. Scharn, O. Svensson, M. Töpel, M. Unterseher, C. Visagie, C. Wurzbacher, A. F. S. Taylor, U. Kõljalg, L. Schriml, and R. H. Nilsson. 2016. Annotating public fungal ITS sequences from the built environment according to the MIxS-Built Environment standard: A report from a May 23–24, 2016 workshop (Gothenburg, Sweden). MycoKeys 16:1-15.
Abrams, P. A. 2009. When does greater mortality increase population size? The long history and diverse mechanisms underlying the hydra effect. Ecology Letters 12:462-474.
Adams, R. I., A. C. Bateman, H. M. Bik, and J. F. Meadow. 2015. Microbiota of the indoor environment: A meta-analysis. Microbiome 3:49. https://microbiomejournal.biomedcentral.com/articles/10.1186/s40168-015-0108-3 (accessed July 14, 2017).
Adams, R. I., S. Bhangar, K. C. Dannemiller, J. A. Eisen, N. Fierer, J. A. Gilbert, J. L. Green, L. C. Marr, S. L. Miller, J. A. Siegel, B. Stephens, M. S. Waring, and K. Bibby. 2016. Ten questions concerning the microbiomes of buildings. Building and Environment 109:224-234.
Adan, O. C. G., and R. A. Samson. 2011. Fundamentals of mold growth in indoor environments and strategies for healthy living. Wageningen, The Netherlands: Wageningen Academic Publishers.
Afgan, E., D. Baker, M. van den Beek, D. Blankenberg, D. Bouvier, M. Čech, J. Chilton, D. Clements, N. Coraor, C. Eberhard, B. Grüning, A. Guerler, J. Hillman-Jackson, G. Von Kuster, E. Rasche, N. Soranzo, N. Turaga, J. Taylor, A. Nekrutenko, and J. Goecks. 2016. The Galaxy platform for accessible, reproducible and collaborative biomedical analyses: 2016 update. Nucleic Acids Research 44(W1):W3-W10.
Ali, A. S., Z. Zanzinger, D. Debose, and B. Stephens. 2016. Open Source Building Science Sensors (OSBSS): A low-cost Arduino-based platform for long-term indoor environmental data collection. Building and Environment 100:114-126.
Araki, A., Y. Eitaki, T. Kawai, A. Kanazawa, M. Takeda, and R. Kishi. 2009. Diffusive sampling and measurement of microbial volatile organic compounds in indoor air. Indoor Air 19(5):421-432.
ASHRAE (American Society of Heating, Refrigerating and Air-Conditioning Engineers). 2012. ASHRAE, ANSI/ASHRAE/ACCA Standard 180-2012. Standard practice for inspection and maintenance of commercial building HVAC systems. Atlanta, GA: American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.
ASHRAE. 2016. ANSI/ASHRAE Standard 62.1: Ventilation for acceptable indoor air quality. Atlanta, GA: ASHRAE.
Bartram, A. K., M. D. Lynch, J. C. Stearns, G. Moreno-Hagelsieb, and J. D. Neufeld. 2011. Generation of multimillion-sequence 16S rRNA gene libraries from complex microbial communities by assembling paired-end illumina reads. Applied and Environmental Microbiology 77(11):3846-3852.
Be, N. A., J. B. Thissen, V. Y. Fofanov, J. E. Allen, M. Rojas, G. Golovko, Y. Fofanov, H. Koshinsky, and C. J. Jaing. 2015. Metagenomic analysis of the airborne environment in urban spaces. Microbial Ecology 69(2):346-355.
Berkeley Lab. 2015. EcoFAB: Initiative researchers will model ecosystem biology on the bench. BioSciences. September 25. http://biosciences.lbl.gov/2015/09/25/ecofab-initiative-researchers-will-model-ecosystem-biology-on-the-bench (accessed July 16, 2017).
Bouslimani, A., C. Porto, C. M. Rath, M. Wang, Y. Guo, A. Gonzalez, D. Berg-Lyon, G. Ackermann, G. J. Moeller Christensen, T. Nakatsuji, L. Zhang, A. W. Borkowski, M. J. Meehan, K. Dorrestein, R. L. Gallo, N. Bandeira, R. Knight, T. Alexandrov, and P. C. Dorrestein. 2015. Molecular cartography of the human skin surface in 3D. Proceedings of the National Academy of Sciences of the United States of America 112(17):E2120-E2129.
Chase, J., J. Fouquier, M. Zare, D. L. Sonderegger, R. Knight, S. T. Kelley, J. Siegel, and J. G. Caporaso. 2016. Geography and location are the primary drivers of office microbiome composition. mSystems1(2):e00022-16. doi:10.1128/mSystems.00022-16.
Coil, D. A., R. Y. Neches, J. M. Lang, W. E. Brown, M. Severance, D. Cavalier, and J. A. Eisen. 2016. Growth of 48 built environment bacterial isolates on board the International Space Station (ISS). PeerJ 4:e1842.
Dedesko, S., and J. A. Siegel. 2015. Moisture parameters and fungal communities associated with gypsum drywall in buildings. Microbiome 3(1):71.
Dorrestein, P. 2016. Mass spectrometry-based visualization of molecules associated with human habits. Presentation to the Committee on Microbiomes of the Built Environment: From Research to Application, October 17.
Dorrestein, P. 2017. Digitizing the chemistry associated with microbes: Importance, current status, and opportunities. In The Chemistry of Microbiomes: Proceedings of a Seminar Series. Washington, DC: The National Academies Press.
EPA (U.S. Environmental Protection Agency). 2006. Building Assessment Survey and Evaluation (BASE) study data on indoor air quality in public and commercial buildings. 402-C-06-002. Washington, DC: EPA.
Flannigan, B., R. A. Samson, and J. D. Miller. 2011. Microorganisms in home and indoor work environments: Diversity, health impacts, investigation and control (2nd ed.). Boca Raton, FL: CRC Press.
Fujimura, K. E., T. Demoor, M. Rauch, A. A. Faruqi, S. Jang, and C. C. Johnson. 2014. House dust exposure mediates gut microbiome Lactobacillus enrichment and airway immune defense against allergens and virus infection. Proceedings of the National Academy of Sciences of the United States of America 111(2):805-810.
Gibbons, S. 2016. The built environment is a microbial wasteland. mSystems 1(2):e00033-16. doi:10.1128/mSystems.00033-16.
Glass, E. M., Y. Dribinsky, P. Yilmaz, H. Levin, R. Van Pelt, D. Wendel, A. Wilke, J. A. Eisen, S. Huse, A. Shipanova, M. Sogin, J. Stajich, R. Knight, F. Meyer, and L. M Schriml. 2014. MIxS-BE: A MIxS extension defining a minimum information standard for sequence data from the built environment. The ISME Journal 8:1-3. http://www.nature.com/ismej/journal/v8/n1/full/ismej2013176a.html (accessed July 16, 2017).
Greenblum, S., R. Carr, and E. Borenstein. 2015. Extensive strain-level copy-number variation across human gut microbiome species. Cell 160(4):583-594.
Gutarowska, B., S. Celikkol-Aydin, V. Bonifay, A. Otlewska, E. Aydin, A. L. Oldham, J. I. Brauer, K. E. Duncan, J. Adamiak, J. A. Sunner, I. B. Beech. 2015. Metabolomic and high-throughput sequencing analysis—modern approach for the assessment of biodeterioration of materials from historic buildings. Frontiers in Microbiology 6:979.
Hanson, C. A., J. A. Fuhrman, M. C. Horner-Devine, and J. B. H. Martiny. 2012. Beyond biogeographic patterns: Processes shaping the microbial landscape. Nature Reviews Microbiology 10(7):497-506.
Harriman, L. G., and J. W. Lstiburek. 2009. The ASHRAE guide for buildings in hot and humid climates (2nd ed.). Atlanta, GA: ASHRAE.
Hoisington, A., J. P. Maestre, J. A. Siegel, and K. A. Kinney. 2014. Exploring the microbiome of the built environment: A primer on four biological methods available to building professionals. HVAC&R Research 20(1):167-175.
Jansson, J. K., and E. S. Baker. 2016. A multi-omic future for microbiome studies. Nature Microbiology 1(5):16049.
Jovel, J., J. Patterson, W. Wang, N. Hotte, S. O’Keefe, T. Mitchel, T. Perry, D. Kao, A. L. Mason, K. L. Madsen, and G. K. S. Wong. 2016. Characterization of the gut microbiome using 16S or shotgun metagenomics. Frontiers in Microbiology 7:459.
Kahlor, L. 2016. Risk communication and information seeking. Presentation to the Committee on Microbiomes of the Built Environment: From Research to Application, June 20.
Kauserud, H., S. Kumar, A. K. Brysting, J. Norden, and T. Carlsen. 2012. High consistency between replicate 454 pyrosequencing analyses of ectomycorrhizal plant root samples. Mycorrhiza 22(4):309-315.
Lauber, C. L., N. Zhou, J. I. Gordon, R. Knight, and N. Fierer. 2010. Effect of storage conditions on the assessment of bacterial community structure in soil and human-associated samples. FEMS Microbiology Letters 307(1):80-86.
Lax, S., D. P. Smith, J. Hampton-Marcell, S. M. Owens, K. M. Handley, N. M. Scott, S. M. Gibbons, P. Larsen, B. D. Shogan, S. Weiss, J. L. Metcalf, L. K. Ursell, Y. Vazquez-Baeza, W. Van Treuren, N. A. Hasan, M. K. Gibson, R. Colwell, G. Dantas, R. Knight, and J. A. Gilbert. 2014. Longitudinal analysis of microbial interaction between humans and the indoor environment. Science 345(6200):1048-1052.
Lax, S., N. Sangwan, D. Smith, P. Larsen, K. M. Handley, M. Richardson, K. Guyton, M. Krezalek, B. D. Shogan, J. Defazio, I. Flemming, B. Shakhsheer, S. Weber, E. Landon, S. Garcia-Houchins, J. Siegel, J. Alverdy, R. Knight, B. Stephens, and J. A. Gilbert. 2017. Bacterial colonization and succession in a newly opened hospital. Science Translational Medicine 9(391):eaah6500. http://stm.sciencemag.org/content/9/391/eaah6500 (accessed July 16, 2017).
Leinonen, R., H. Sugawara, and M. Shumway. 2011. The sequence read archive. Nucleic Acids Research 39:D19-D21.
Lemos, L. N., R. R. Fulthorpe, and L. F. Roesch. 2012. Low sequencing efforts bias analyses of shared taxa in microbial communities. Folia Microbiologica 57(5):409-413.
Lindgreen, S., K. L. Adair, and P. P. Gardner. 2016. An evaluation of the accuracy and speed of metagenome analysis tools. Scientific Reports 6:19233.
Macher, J. M., J. Douwes, B. Prezant, and T. Reponen. 2013. Bioaerosols. In Aerosols handbook: Measurement, dosimetry, and health effects, 2nd ed., edited by L. S. Ruzer and N. H. Harley. Boca Raton, FL: CRC Press. Pp. 285-344.
McKain, N., B. Genc, T. J. Snelling, and R. J. Wallace. 2013. Differential recovery of bacterial and archaeal 16S rRNA genes from ruminal digesta in response to glycerol as cryoprotectant. Journal of Microbiological Methods 95(3):381-383.
Merchant, N., E. Lyons, S. Goff, M. Vaughn, D. Ware, D. Micklos, and P. Antin. 2016. The iPlant collaborative: Cyberinfrastructure for enabling data to discovery for the life sciences. PLOS Biology 14(1):e1002342.
Moschandreas, D. J., K. R. Pagilla, and L. V. Storino. 2008. Time and space uniformity of indoor bacteria concentrations in Chicago area residences. Aerosol Science and Technology 37:899-906.
Mudarri, D. H. 1997. Potential correction factors for interpreting CO2 measurements in buildings. ASHRAE Transactions 103(2):244-255.
NASEM (National Academies of Sciences, Engineering, and Medicine). 2017a. Communicating science effectively: A research agenda. Washington DC: The National Academies Press.
NASEM. 2017b. The value of social, behavioral, and economic sciences to national priorities: A report for the National Science Foundation. Washington DC: The National Academies Press.
Nayfach, S., and K. S. Pollard. 2016. Toward accurate and quantitative comparative metagenomics. Cell 166(5):1103-1116.
Nayfach, S., B. Rodriguez-Mueller, and K. S. Pollard. 2016. An integrated metagenomics pipeline for strain profiling reveals novel patterns of bacterial transmission and biogeography. Genome Research 26(11):1612-1625.
NIST (National Institute of Standards and Technology). 2012. Challenges in microbial sampling in indoor environments: Workshop report summary. NIST Technical Note 1737. http://www.microbe.net/wp-content/uploads/2012/03/Sloan-Roport-Final-TN-1737.pdf (accessed July 16, 2017).
NRC (National Research Council) and IOM (Institute of Medicine). 2015. BioWatch PCR assays: Building confidence, ensuring reliability. Washington, DC: The National Academies Press.
Peccia, J., and M. Hernandez. 2006. Incorporating polymerase chain reaction-based identification, population characterization, and quantification of microorganisms into aerosol science: A review. Atmospheric Environment 40:3941-3961.
Persily, A. K. 1997. Evaluating building IAQ and ventilation with carbon dioxide. ASHRAE Transactions 103(2):193-204.
Pinto, A. J., and L. Raskin. 2012. PCR biases distort bacterial and archaeal community structure in pyrosequencing datasets. PLOS ONE 7(8):e43093.
Prussin, A. J., and L. C. Marr. 2015. Sources of airborne microorganisms in the built environment. Microbiome 3(1):78.
Qian, J., D. Hospodsky, N. Yamamoto, W. W. Nazaroff, and J. Peccia. 2012. Size-resolved emission rates of airborne bacteria and fungi in an occupied classroom. Indoor Air 22(4):339-351.
Qian, J., J. Peccia, and A. R. Ferro. 2014. Walking-induced particle resuspension in indoor environments. Atmospheric Environment 89:464-481.
Quinn, R. A., J. A. Navas-Molina, E. R. Hyde, S. J. Song, Y. Vázquez-Baeza, G. Humphrey, J. Gaffney, J. J. Minich, A. V. Melnik, J. Herschend, J. DeReus, A. Durant, R. J. Dutton, M. Khosroheidari, C. Green, R. da Silva, P. C. Dorrestein, and R. Knight. 2016. From sample to multi-omics conclusions in under 48 hours. mSystems 1(2):e00038-16. doi:10.1128/ mSystems.00038-16.
Ramos, T., and B. Stephens. 2014. Tools to improve built environment data collection for indoor microbial ecology investigations. Building and Environment 81:243-257.
Ramos, T., S. Dedesko, J. A. Siegel, J. A. Gilbert, and B. Stephens. 2015. Spatial and temporal variations in indoor environmental conditions, human occupancy, and operational characteristics in a new hospital building. PLOS ONE 10(3):e0118207.
Randle-Boggis, R. J., T. Helgason, M. Sapp, and P. D. Ashton. 2016. Evaluating techniques for metagenome annotation using simulated sequence data. FEMS Microbiology Ecology 92(7). doi:10.1093/femsec/fiw095.
Rhee, S.-K., X. Liu, L. Wu, S. C. Chong, X. Wan, and J. Zhou. 2004. Detection of genes involved in biodegradation and biotransformation in microbial communities by using 50-mer oligonucleotide microarrays. Applied and Environmental Microbiology 70(7):4303-4317.
Saraf, A., L. Larsson, H. Burge, and D. Milton. 1997. Quantification of ergosterol and 3-hydroxy fatty acids in settled house dust by gas chromatography-mass spectrometry: Comparison with fungal culture and determination of endotoxin by a Limulus amebocyte lysate assay. Applied Environmental Microbiology 63(7):2554-2559.
Schloss, P. D., D. Gevers, and S. L. Westcott. 2011. Reducing the effects of PCR amplification and sequencing artifacts on 16S rRNA-based studies. PLOS ONE 6(12):e27310.
Shogan, B. D., D. P. Smith, A. I. Packman, S. T. Kelley, E. M. Landon, S. Bhangar, G. J. Vora, R. M. Jones, K. Keegan, B. Stephens, T. Ramos, B. C. Kirkup, Jr., H. Levin, M. Rosenthal, B. Foxman, E. B. Chang, J. Siegel, S. Cobey, G. An, J. C. Alverdy, P. J. Olsiewski, M. O. Martin, R. Marrs, M. Hernandez, S. Christley, M. Morowitz, S. Weber, and J. Gilbert. 2013. The Hospital Microbiome Project: Meeting report for the 2nd Hospital Microbiome Project, Chicago, USA, January 15, 2013. Standards in Genomic Sciences 8(3):571-579.
Singer, E., B. Andreopoulos, R. M. Bowers, J. Lee, S. Deshpande, J. Chiniquy, D. Ciobanu, H.-P. Klenk, M. Zane, C. Daum, C., A. Clum, J.-F. Cheng, A. Copeland, and T. Woyke. 2016. Next generation sequencing data of a defined microbial mock community. Scientific Data 3:160081.
Smith, D., J. Alverdy, G. An, M. Coleman, S. Garcia-Houchins, J. Green, K. Keegan, S. T. Kelley, B. C. Kirkup, L. Kociolek, H. Levin, E. Landon, P. Olsiewski, R. Knight, J. Siegel, S. Weber, and J. Gilbert. 2013. The Hospital Microbiome Project: Meeting report for the 1st Hospital Microbiome Project Workshop on sampling design and building science measurements, Chicago, USA, June 7–8, 2012. Standards in Genomic Sciences 8(1):112-117.
Stein, M. M., C. L. Hrusch, J. Gozdz, C. Igartua, V. Pivniouk, S. E. Murray, N., J. G. Ledford, M. Marques dos Santos, R. L. Anderson, N. Metwali, J. W. Neilson, R. M. Maier, J. A. Gilbert, M. Holbreich, P. S. Thorne, F. D. Martinez, E. von Mutius, D. Vercelli, C. Ober, and A. I. Sperling. 2016. Innate immunity and asthma risk in Amish and Hutterite farm children. New England Journal of Medicine 375(5):411-421.
Stephens, B. 2016. What have we learned about the microbiomes of indoor environments? mSystems 1(4):e00083-e00116. doi:10.1128/mSystems.00083-16.
Taylor, D. L., W. A. Walters, N. J. Lennon, J. Bochicchio, A. Krohn, J. G. Caporaso, and T. Pennanen. 2016. Accurate estimation of fungal diversity and abundance through improved lineage-specific primers optimized for illumina amplicon sequencing. Applied and Environmental Microbiology 82(24):7217-7226.
Uyaguari-Diaz, M. I., M. Chan, B. L. Chaban, M. A. Croxen, J. F. Finke, J. E. Hill, M. A. Peabody, T. Van Rossum, C. A. Suttle, F. S. L. Brinkman, J. Isaac-Renton, N. A. Prystajecky, and P. Tang. 2016. A comprehensive method for amplicon-based and metagenomic characterization of viruses, bacteria, and eukaryotes in freshwater samples. Microbiome 4(1):20.
Vizcaíno, J. A., A. Csordas, N. del-Toro, J. A. Dianes, J. Griss, I. Lavidas, G. Mayer, Y. Perez-Riverol, Y., F. Reisinger, T. Ternent, Q.-W. Xu, R. Wang, and H. Hermjakob. 2016. 2016 update of the PRIDE database and its related tools. Nucleic Acids Research 44(22):D447-D456.
Wang, M., J. J. Carver, V. V. Phelan, L. M. Sanchez, N. Garg, Y. Peng, D. D. Nguyen, J. Watrous, C. A. Kapono, T. Luzzatto-Knaan, C. Porto, A. Bouslimani, A. V. Melnik, M. J. Meehan, W. T. Liu, M. Crüsemann, P. D. Boudreau, E. Esquenazi, M. SandovalCalderón, R. D. Kersten, L. A. Pace, R. A. Quinn, K. R. Duncan, C. C. Hsu, D. J. Floros, R. G. Gavilan, K. Kleigrewe, T. Northen, R. J. Dutton, D. Parrot, E. E. Carlson, B. Aigle, C. F. Michelsen, L. Jelsbak, C. Sohlenkamp, P. Pevzner, A. Edlund, J. McLean, J. Piel, B. T. Murphy, L. Gerwick, C. C. Liaw, Y. L. Yang, H. U. Humpf, M. Maansson, R. A. Keyzers, A. C. Sims, A. R. Johnson, A. M. Sidebottom, B. E. Sedio, A. Klitgaard, C. B. Larson, P. C. A. Boya, D. Torres-Mendoza, D. J. Gonzalez, D. B. Silva, L. M. Marques, D. P. Demarque, E. Pociute, E. C. O’Neill, E. Briand, E. J. Helfrich, E. A. Granatosky, E. Glukhov, F. Ryffel, H. Houson, H. Mohimani, J. J. Kharbush, Y. Zeng, J. A. Vorholt, K. L. Kurita, P. Charusanti, K. L. McPhail, K. F. Nielsen, L. Vuong, M. Elfeki, M. F. Traxler, N. Engene, N. Koyama, O. B. Vining, R. Baric, R. R. Silva, S. J. Mascuch, S. Tomasi, S. Jenkins, V. Macherla, T. Hoffman, V. Agarwal, P. G. Williams, J. Dai, R. Neupane, J. Gurr, A. M. Rodríguez, A. Lamsa, C. Zhang, K. Dorrestein, B. M. Duggan, J. Almaliti, P. M. Allard, P. Phapale, L. F. Nothias, T. Alexandrov, M. Litaudon, J. L. Wolfender, J. E. Kyle, T. O. Metz, T. Peryea, D. T. Nguyen, D. VanLeer, P. Shinn, A. Jadhav, R. Müller, K. M. Waters, W. Shi, X. Liu, L. Zhang, R. Knight, P. R. Jensen, B. Ø. Palsson, K. Pogliano, R. G. Linington, M. Gutiérrez, N. P. Lopes, W. H. Gerwick, B. S. Moore, P. C. Dorrestein, and N. Bandeira. 2016. Sharing and community curation of mass spectrometry data with Global Natural Products Social Molecular Networking. Nature Biotechnology 34(8):828-837.
Weiss, S., W. Van Treuren, C. Lozupone, K. Faust, J. Friedman, Y. Deng, L. C. Xia, Z. Z. Xu, L. Ursell, E. J. Alm, A. Birmingham, J. A. Cram, J. A. Fuhrman, J. Raes, F. Sun, J. Zhou, and R. Knight. 2016. Correlation detection strategies in microbial data sets vary widely in sensitivity and precision. The ISME Journal 10(7):1669-1681.
Wu, L., D. K. Thompson, G. Li, R. A. Hurt, J. M. Tiedje, and J. Zhou. 2001. Development and evaluation of functional gene arrays for detection of selected genes in the environment. Applied and Environmental Microbiology 67(12):5780-5790.
Wu, L., X. Liu, C. W. Schadt, and J. Zhou. 2006. Microarray-based analysis of subnanogram quantities of microbial community DNAs by using whole-community genome amplification. Applied and Environmental Microbiology 72(7):4931-4941.
Yao, M., T. Zhu, K. Li, S. Dong, Y. Wu, X. Qiu, B. Jiang, L. Chen, and S. Zhen. 2009. Onsite infectious agents and toxins monitoring in 12 May Sichuan earthquake affected areas. Journal of Environmental Monitoring 11(11):1993-2001.
Yooseph, S., C. Andrews-Pfannkoch, A. Tenney, J. McQuaid. S. Williamson, M. Thiagarajan, D. Brami, L. Zeigler-Allen, J. Hoffman, J. B. Goll, D. Fadrosh, J. Glass, M. D. Adams, R. Friedman, and J. C. Venter. 2013. A metagenomics framework for the study of airborne microbial communities. PLOS ONE 8(12):e81862.
Zhou, J. Z., L. Wu, Y. Deng, X. Zhi, Y.-H. Jiang, Q. Tu, J. Xie, J. D. Van Nostrand, Z. He, and Y. Yang. 2011. Reproducibility and quantitation of amplicon sequencing-based detection. The ISME Journal 5(8):1303-1313.
Zhou, J. Z., Z. He, Y. Yang, Y. Deng, S. G. Tringe, and L. Alvarez-Cohen. 2015. High-throughput metagenomic technologies for complex microbial community analysis: Open and closed formats. mBio 6(1). doi:10.1128/mBio.02288-14.
Zou, H., Z. Chen, H. Jiang, L. Xie, and C. Spanos. 2017. Accurate indoor localization and tracking using mobile phone inertial sensors, WiFi and iBeacon. In 2017 IEEE International Symposium on Inertial Sensors and Systems (INERTIAL), Kauai, HI. Pp. 1-4. doi:10.1109/ISISS.2017.7935650.
This page intentionally left blank.