April 15, 2014 (Vol. 34, No. 8)

Richard A. A. Stein M.D., Ph.D.

The bacteria inhabiting our bodies greatly outnumber our own cells, by approximately 10 times, illustrating that we are, in a sense, more bacterial than human.

An alternative (but still disconcerting) view is that we represent just one of many ecosystems dominated by microorganisms, whose versatility and adaptability has allowed them to become the dominant form of life in nearly every ecosystem on the planet.

While pervasive, bacteria remain secretive. Our ability to interrogate microbial communities has been severely limited because the vast majority of bacterial species fail to grow under laboratory-established conditions. Only an overrepresented minority of culturable bacteria has suffered our scrutiny.

“There is a lot of hidden microbial diversity that we have been ignoring,” says Jonathan A. Eisen, Ph.D., a professor with dual appointments (evolution and ecology; medical microbiology and immunology) at the University of California, Davis. Advances in DNA sequencing technologies have greatly improved our ability to characterize unculturable microorganisms from microbial communities. These technologies are being used in several research efforts in Dr. Eisen’s group, helping it examine the contributions of both culturable and also unculturable microorganisms to microbial diversity.

“We are using a combination of targeted surveys of ribosomal RNA genes and random surveys of “metagenomes” of whole communities,” notes Dr. Eisen. “As sequencing has gotten cheaper, we have been able to devote a higher fraction of our work toward metagenomics.” To a great extent, decisions on the specific technical approach are shaped by the particular question being asked about a particular biological system. “One of the reasons we still do a lot of sequence-based studies of microbial communities is that these do not have the biases that culturing brings in.”

Researchers at the University of California, Davis, compared SFams and noticed a network of relationships between families. The network’s components tend to have a minimally connected topology, as illustrated by this random sampling of 250 components from the network. Dots and lines represent relationships between SFams. Most components are populated by a small number of SFams. This image appeared in Sharpton et al. BMC Bioinformatics 2012 13: 264.

Storage, Processing, Analysis

With the developments in sequencing platforms that opened the possibility to generate reads in a very cost-effective manner, the challenge has shifted from performing the sequencing to storing and processing the vast datasets. “Even as sequencing gets cheaper and cheaper, there will likely always be a balance between rRNA surveys and metagenomics—for a set amount of money, one can do more rRNA surveys than metagenomic analyses,” asserts Dr. Eisen.

To assist the phylogenetic analysis of unculturable microbial species, Dr. Eisen and colleagues developed an open-source pipeline to analyze metagenomic datasets. “PhyloSift, our newest software, allows one to carry out phylogeny-driven analyses of metagenomic datasets in much the same way many have been analyzing rRNA datasets for years,” remarks Dr. Eisen.

These advances are opening a new era in microbial ecology, which has a history of defying experimental inquiry. “The microbial ecosystem is different in different locations of the human body,” observes Dr. Eisen.  “Maybe we should view an individual person the way we view a planet, which has different organisms present in the savannah as compared to the rainforest, for example.”

In one of its studies, Dr. Eisen’s group used metabolomic profiling to survey the microbiota in patients with small bowel transplants. The investigators found an enrichment of metabolites associated with aerobic respiration in patients undergoing ileostomy. This result indicated that ileostomy profoundly alters the composition of the intestinal microbial flora, supporting the hypothesis that the microbial flora in the ileum exists in two alternate and stable states.

Host-Pathogen Interactions

“If we look back at the way we learned about host-pathogen interactions, this provides a good example of the path that we most likely will follow to explore the metagenome,” says Andrew K. Benson, Ph.D., professor of biotechnology at the University of Nebraska, Lincoln. In the early days of microbiology, much of the knowledge about the host-pathogen interface emerged from reductionist models, which assumed that virulence is shaped by either the host or the pathogen.

“Nobody would have imagined that the host-pathogen interaction is as sophisticated and as ornate as we now know,” comments Dr. Benson. “With the metagenome on the same track, we will be amazed at how elaborate microbial systems are, only that this time it will be on a completely different level.”

Building on previous work that explored the genetics of pathogen evolution, Dr. Benson’s group is focusing on our understanding of the interactions between the host and the microbiome. The group is intent on testing the hypothesis that it should be possible to detect, at least partly, the contributions that host genetic contributions have on the microbiome.

Initial studies on mouse models have shown that host genetics shapes the composition of the gut microbiome. “As we became more sophisticated, we incorporated diet into the equation,” explains Dr. Benson. “We are now able to test, at the metagenome level, whether the host shapes only the species composition or whether it independently also shapes the functional content of the microbiome.”

Some of the challenges in addressing this aspect are intricately linked to sequencing, because switching from 16S ribosomal RNA-based sequencing to full metagenome sequencing presents several of the difficulties that are known from other omics disciplines.

“Our quantitative genetics studies typically involve hundreds of thousands of samples and a scale that is usually an order of magnitude over what a lot of other projects involve,” adds Dr. Benson. “This adds a whole other element of difficulty.”

Dr. Benson and colleagues recently established Metagenome Analytics, a company that brings together investigators with broad expertise in sequencing analyses and food sciences. It is trying to use metagenomics to bridge the gap between service providers and users in the food industry. Specifically, the company intends to help users with sequencing, data analysis, and the actionable interpretation of results.

“These activities,” declares Dr. Benson, “will [encompass] the identification of organisms or populations of interest, the identification of markers for organisms of interest, traceback analyses, the development of rapid and inexpensive assays, and the development of action levels.”

Microbial Gene Exchange

“Our analysis of genome and metagenome data revealed large regions of identical sequencing even between the genomes of organisms that otherwise seem pretty distant from each other,” says Susannah G. Tringe, Ph.D., head of the metagenome program at the Department of Energy’s Joint Genome Institute. This remark references Dr. Tringe’s collaboration with a group of investigators led by Ricardo Cavicchioli, Ph.D., a professor at the School of Biotechnology and Biomolecular Sciences at the University of New South Wales.

Dr. Tringe and colleagues conducted a metagenomic analysis on samples from the Deep Lake in Antarctica. This isolated system is unique. Not only is it hypersaline, it also remains liquid at temperatures as low as −20°C (−4°F).

Analysis of the metagenomic data and studies on the genomes of four isolates unveiled a remarkable degree of gene exchange across several strains, even though they represented distinct and distantly related haloarcheal species. This process involved the exchange of chromosomal regions up to 35 kb in size, and sharing almost 100% identity.

“Most of this study relied heavily on computational analyses,” notes Dr. Tringe. “The results suggest that something between the genomes was swapped in the recent past, as opposed to millions of years ago.”

Researchers representing the Department of Energy’s Joint Genome Institute conducted a metagenomics analysis on samples taken in 2008 from Antarctica’s Deep Lake, a unique and isolated (and extreme) environment. The researchers found that the microbial neighborhood at Deep Lake is fairly homogenous. This finding runs counter to the usual expectation that gene exchange across species boundaries should occur infrequently. Indeed, it appears to occur comparatively frequently among the haloarchaea living in Deep Lake’s hypersaline environment.

Community Dynamics

 “We apply sequencing and bioinformatics to address and understand how microbes interact in ecosystems to form structured communities and how, in those communities, they affect changes in the environment,” says Jack A. Gilbert, Ph.D., an associate professor of ecology and evolution at the University of Chicago and an environmental microbiologist at the Argonne National Laboratory. Dr. Gilbert and colleagues are involved in several large-scale metagenomics research initiatives.

One of these initiatives, the Earth Microbiome Project, is a multidisciplinary endeavor that proposes to analyze 100,000–200,000 natural samples to characterize microbial communities across the globe from a broad range of environments. This would help establish a catalog of the microbial community structure from different environments, and use the respective databases to address fundamental questions about the ecology of life. “We collect samples originating from many sources, such as humans, marine water, and sediments, without discriminating, and characterize the microbiome structure across the planet,” explains Dr. Gilbert.

Another project, focusing on the hospital microbiome, proposes to examine more than 15,000 hospital samples collected over two or three years. The goal is to provide a picture of patients’ microbial colonization in hospitals. With 5% of the patients admitted to U.S. medical facilities developing hospital-acquired infections, many of them caused by strains resistant to antimicrobials, this endeavor has major implications for public health.

Yet another project, the Chicago Waterways Metagenomic Initiative, involves the collection of samples from multiple locations around the city of Chicago every month over several years. The idea is to build a comprehensive picture of the dynamic impact that the city architecture exerts on microbial communities.

According to Dr. Gilbert, all of these projects share a key challenge: “To really have a translational impact, we need to understand how to quantitate changes in the abundance of key organisms and their functional impact on the community. “We are still struggling with that.”

Snapshot of a taxonomic sharing network—the large dots are samples (red: dog skin; yellow: human skin; green: home surfaces) whereby proximity between them suggests a more similar microbiome. The small yellow dots are bacterial taxa, and the edges are colored according to the environments in which these taxa are found. [Jack Gilbert and Simon Lax, University of Chicago]

Visualizing Compositional Changes

“Historically, we resorted to using very simplistic views,” says Mitchell L. Sogin, Ph.D., professor of molecular biology, cell biology, and biochemistry at the Marine Biological Laboratory. “Microbial communities, however, are very complex, and there are different levels in terms of visualizing the community composition.”

Dr. Sogin’s group recently created VAMPS (Visualization and Analysis of Microbial Population Structures), a free, web-based service that allows investigators to interactively compare sequencing datasets and generate and test hypotheses. “We intended this to be a database that is easy to use, so people do not need to know about programs and scripts,” remarks Dr. Sogin.

In addition to allowing users to capture the diversity of microbial communities and the relationship between communities, VAMPS also encourages users to share data and facilitates the identification of shared patterns.

Research in Dr. Sogin’s lab is focused, in part, on measuring fecal pollution to identify human sewage that impacts bodies of water. Historically, investigators have relied on the presence of indicator organisms that are not necessarily pathogenic, such as E. coli, to determine whether pollution has occurred. “The difficulty is that E. coli is a very nonspecific indicator,” says Dr. Sogin.
As an alternative, microbial communities that are present in a particular sample can be compared to a community of samples of human or animal origin. This approach, notes Dr. Sogin, can help investigators characterize specific signatures. That is, the signatures can provide much more specificity in terms of the source.

In a recent study, Dr. Sogin and colleagues used a microbial signature to examine fecal pollution in 40 sewage samples near Milwaukee, off the urban shore of Lake Michigan, and revealed that an urban fecal footprint extends at least 8 km offshore. However, even when visualizing changes in the composition of microbial populations, understanding the functional impact of the changes is still a significant challenge.

“It may very well be that we need to be looking not so much at the presence, absence, increase, or decrease of a microorganism,” continues Dr. Sogin, “but at the functional repertoire and the functional changes that are represented by the entire community.”

With only a minority of microbial species being currently culturable by conventional methods, metagenomics is poised to profoundly reshape our perspectives on microbial communities and on the microbe-host interface. The developments in this field were catalyzed by transformative moments in several disciplines including omics, biotechnology, and bioinformatics. This opened opportunities to capture, at an unprecedented level of scrutiny, the bacterial world that has orchestrated life, and continues to do so, on the global stage that Stephen Jay Gould so eloquently called the “Planet of the Bacteria.”

Microbial Link between Permafrost Thaws and Methane Emissions

“We are trying to understand which microorganism is primarily responsible for methane gas generation in the thawing permafrost,” says Gene W. Tyson, Ph.D., associate professor and deputy director at the University of Queensland’s Australian Centre for Ecogenomics. Dr. Tyson leads an international team of scientists that is examining permafrost in Stordalen Mire, a region in northern Sweden where the ecosystem has undergone massive changes in recent decades.

From the three habitat strata that were described in this region, intact permafrost receded, while partially thawed and fully thawed permafrost expanded between 1970 and 2000. These changes were accompanied by an increase in methane gas emissions.

By using metagenomics approaches, Dr. Tyson and colleagues revealed that a single archeal phylotype dominates the thawing and thawed regions of the permafrost. The species belongs to an uncultivated lineage, and metagenomic sequencing led to the reconstruction of its genome, which helped identify the highly expressed genes that are involved in methanogenesis.

“In the next few years, new approaches will likely allow us to use metagenomic data, particularly deep metagenomic data, to look at hundreds or thousands of microbial genomes and visualize the community from a different perspective,” explains Dr. Tyson.

“Ultimately, this helps us understand the tree of life.”