WHAT ARE THE FUTURE RESEARCH OPPORTUNITIES AND CHALLENGES IN MICROBIOME RESEARCH?
Can We Identify Cross-System and Cross-Platform Commonalities and Opportunities for Collaboration and Integration?
During the final session of the seminar series, a panel was convened to lead the discussion around cross-system and cross-platform commonalities and opportunities for collaboration and integration. The discussion revealed, first and foremost, how diverse and potentially disconnected the ecosystem-derived research communities are and where bridges might potentially be constructed. By engaging the seminar participants, the panel worked to arrive at several key themes that signal opportunities for or barriers to collaboration. The most important learning from this seminar series, however, was the need for research and funding agencies to continue these discussions to promote and enable trans-system collaborations. This epilogue captures highlights of the panel discussion.
Accurate Genome Annotations
A major limitation for understanding the chemical reactions that microbiomes catalyze and the metabolites they synthesize, regardless of system, is the lack of accurate genome annotations. Still missing is an extensive amount of functional biology encoded within genomes, even in well-studied model systems such as Escherichia coli; there remain many genes of unknown function that, in some cases, are among the most highly expressed. Another problem is that some genes can have multiple functions that cannot be resolved based on transcriptomics alone. Many genomic tools currently exist that can reveal the organisms present within a microbiome, but there is a challenge in integrating different types of data to reveal underlying mechanisms. The lack of a comprehensive understanding of individual organisms genetic complement in communities limits understanding of how individuals collectively work together as functional microbiomes to chemically interact with each other and their environment.
A possibility for advancing the field of microbiome chemistry discussed during the panel discussion is to develop model synthetic or simplified natural microbiomes that can be shared among researchers and characterized
in detail. For example, one current limitation in human gut microbiome research is that there is no common definition of what constitutes a human gut microbiome. In the absence of such a definition, it is difficult to compare, and impossible to standardize results across different laboratories. The tools of systems biology can be used to deconstruct and develop a comprehensive understanding of model communities, including the detailed characterization of genes and how and when metabolites are synthesized. Moreover, the chemical interactions between community members and their habitat or host can be discerned. One attractive concept would be to develop model microbiomes that are broadly accepted and informative across systems. They would need to be highly reproducible and readily manipulated, which in turn would require that the members of such model microbiomes be genetically well characterized. These could be disseminated among different laboratories, whereby the scientific community would work on and share information, data, and computational models. Such systems could be extremely useful for gaining new insights into microbiome chemistries and causal mechanisms, and for testing specific ecological and biogeochemical hypotheses. While such model systems can be highly informative and would be valuable to microbiome researchers, they are not an all-encompassing solution. For example, they may oversimplify highly complex ecosystems found in nature, such as those in the marine environment.
Possible systems to focus on and develop models for would be those where there is already a considerable body of knowledge, such as the human gut, a model plant’s rhizosphere, and marine coral. Another attractive option could be the oligotrophic open ocean, where time series data have been collected and major microbiome players are well known. Ocean microbiomes can also serve as important historical records of what is currently happening in these ecosystems, which can provide a baseline for quantifying environmental change and its impacts. Other useful model systems focus on understanding interactions between animal or plant hosts and their respective microbiomes. For example, researchers can use control of the host to study known beneficial organisms and their reciprocal impacts. Such approaches could provide an outstanding framework for moving forward in microbiome chemistry research and asking important questions, such as how microbiomes may play critical roles in host function; microbiomes may be unseen instigators of interactions between higher organisms. Efforts may include understanding commonalities between microbiomes of different hosts, regions, and identifying whether patterns and similar interactions are occurring that can begin to formulate underlying principles of microbiome function.
Characterizing Secondary Metabolites
While scientists can currently identify about 95% of the primary metabolites produced by microbiomes, only 5%, at best, of all secondary metabolites can be classified; and many primary ions present in mass spectra of microbiome metabolite profiles cannot be identified. For no single organism, not even E. coli, do we know the entire secondary metabolome. This vast array of unknown secondary metabolites has been called “metabolic dark matter” because it remains uncharacterized. These metabolites are extremely important, as microbes use them to communicate, engage in warfare, and to help catalyze critical chemical reactions. In essence, we know only 3-5% of the chemical lexicon of microbiomes. Considerable resources have been expended in sequencing genomes and metagenomes with considerable success. A similar approach could be used to characterize the chemistries of microbiomes, whereby entire microbiome metabolomes could be determined and catalogued for the benefit of the entire community. Envisioned is an effort equivalent to sequencing the human genome, where advanced technologies could be developed and employed to identify all metabolites. Such an endeavor would provide a comprehensive chemical catalog of microbiome metabolites to serve as a resource for the entire community. In addition to identifying metabolites, measuring the abundance of specific proteins may be essential, especially those that are difficult to identify, such as transporters. Also, measuring chemical fluxes within and between cells and in situ reaction rate constants would yield critical information for developing accurate metabolic models. In addition, information on uptake rates and how they change with environmental conditions would advance quantitative models which cannot be developed in the absence of such parameters.
Grounding Our Understanding of Function
Another major challenge is to understand the true function of microbiomes within their native habitat as opposed to inferring function from genome sequencing or investigation using lab-based cultivation outside of their environmental context. While systems that are sufficiently simple to work with can be readily identified and developed as models, they may or may not be representative of in vivo microbiomes. An array of complex processes works collectively in microbiomes, as they do in larger ecosystems. For example, consumer–prey interactions occur simultaneously with carbon and nitrogen processing in complex ways that are not readily reproducible in the laboratory. While model microbiomes are powerful and can provide detailed mechanistic understanding, they may not be environmentally representative, leading to disconnects between what has been learned in model systems and what is taking place in the environment.
Incorporating Knowledge from Experimental Ecologists
Experimental community ecology has developed a detailed understanding of how some ecosystems function, such as forests and corals. Many fundamental processes at play in these ecosystems likely also apply to microbiomes. The microbiological research community has traditionally focused on autecology and has been constrained to working on individual organisms that can be cultivated in the laboratory. Detailed investigations of physiology and genetics have been undertaken in an attempt to use such information to predict how organisms would interact in nature, but it did not work very well. While experimental macroecologists do not have all the same powerful genomic tools that microbiologists have, they have considerable experience with complex systems and how to manipulate them to gain detailed insights into how they function. A merger of many disciplines, beginning with microbial and macroecologies in concert with genomics, chemical ecology, and other fields would help to gain a more comprehensive understanding of the functioning of microbiomes and the chemicals they produce and the reactions they catalyze.
Understanding complex microbiome behavior, such as how microbes sense and respond to each other and the conditional nature of those interactions, is important and can help scientists link knowledge about chemistry and metabolomics with microbial taxonomic networks. Organisms that are rare and chemicals that are present in exceedingly low concentrations can have extremely important impacts on microbiome functions.
Understanding the chemistry of complex microbiomes may also be achieved by taking more of a traditional engineering approach. In many cases, it will not be practical or even possible to understand all the details of how complex microbiomes, such as those that exist in soil, function. Alternatively, these systems can be teased apart into fundamental components and studied from the bottom up using individual components. Using this approach, complex interactions are reduced or ignored, distilling the system into simpler components. Experimental approaches coupled with models can be used to identify simpler interactions between two organisms or even two genes. A phenomenological approach can be used, whereby competition can be tested in a combinatorial manner where the outcomes can be measured and compared with models. A primary goal would be to gain a basic understanding of design rules for chemical interactions between organisms in microbiomes. Computational models, such as community flux balance, are critical for understanding, predicting, and engineering complex microbiomes, but such efforts are nascent. New approaches that rapidly develop discoveries and new knowledge into models can bridge current knowledge gaps using computation biology.
Emerging Developments in Research Capabilities
A commonality across different systems is the technical capabilities—including genomics, transcriptomics, proteomics, and metabolomics—being used in research; this commonality has generated considerable synergy across systems. One important consideration with such techniques is that microbes, due to their small size and high
surface-to-volume ratio, are highly dependent on local chemical and physical conditions. Nearly all -omics methods require hundreds to millions of cells to make measurements of genes, transcripts, proteins, and metabolites; hence, values obtained represent population averages. We know that there is often tremendous heterogeneity, even within a single clonal microbial population growing under steady-state and well-mixed conditions; such heterogeneity is even greater in biofilms. Therefore, new sensitive and high-throughput single-cell techniques, like chemical imaging, are needed to perform measurements at fine spatial and temporal resolutions. In concert, new computational tools are needed to analyze these data in a statistically robust manner. Because of their small size, microorganisms live within and respond to fine-scale chemical and physical gradients in pH, dissolved ions, nutrients, and light that greatly influence the behavior of individual cells. By simultaneously measuring fine-scale, dynamic chemical gradients—in which microbial communities reside and interact—with single cell measurements, we may be able to analyze the interactions between organisms at the scale that the interactions occur. The scale at which we currently probe the interactions between organisms is not at the scale interactions occur. Assuming such capabilities can be developed, it is important that they are accessible community-wide to ensure the broadest impact.
Understanding the chemistry of microbiomes is simply too large of a challenge for any single funding institution; it will take a large cooperative, collective effort to provide understanding beyond the present descriptive phase to a more mechanistic phase. Collaboration—not only among scientists, but among funding agencies—was noted by many participants throughout the open discussion for transforming microbiome research from an observational and descriptive science to one that is predictive and based on known mechanisms and principles. Common funding sources for tools and technologies, including system agnostic technologies, can bring communities together. For example, a multiagency microbiome-enabling technology development program that is system agnostic could be developed. Within 5-10 years, the community needs to have in hand a set of simplifying principles, inside a theoretical framework, for the assembly and function of microbiomes. Future workshops to follow on and more comprehensively discuss the concepts and ideas highlighted in this seminar were encouraged throughout the seminar series.