Bookmark and Share

Spotlight Live: A Microbial Manifesto (Transcript)

Original webcast is available here.

Microbiome of skinArtistic representation of 3D mapping of the chemistry and microbes of the human skin, based on a study by Pieter Dorrestein and Rob Knight at UC San Diego. (Credit: Theodore Alexandrov)

MICROBES COULD SOON BE at the top of the world's big science list. That's the goal of a consortium of 48 scientists from 50 U.S. institutions who, late last year, proposed the "Unified Microbiome Initiative" – a national effort to advance our understanding of microbiomes, communities of single-celled organisms such as bacteria, viruses and fungi. (Their call was echoed by other scientists from the U.S., Germany and China.)

Ultimately, the researchers hope to harness microbiomes to cure disease, reduce drug resistance, reclaim exhausted farmland, reduce or eliminate agricultural fertilizers and pesticides, and produce useful fuels and chemicals.

Reaching those ambitious goals will require an equally ambitious effort to develop new tools and collaborations for studying microbiomes. Fortunately, scientists will be able to capitalize on a recent revolution in our understanding of the microbial world.

Scientists can now identify microbes by their DNA, and they’ve discovered that microbial communities are far more diverse than anyone ever imagined, including tens or even hundreds of thousands of different microbial species, all interacting with one another. In fact, wherever they have looked, researchers have found microbiomes influencing our world. In the human gut, microbes not only aid digestion, but also impact obesity, allergies and even brain development. Beyond our bodies, they created the Earth's oxygen-rich atmosphere, and enable plant and ocean life to thrive.

But DNA testing cannot explain how microbial genes function and how these organisms work together. Only with that level of understanding, will scientists be able to harness microbiomes to improve human health and the environment.

In January, The Kavli Foundation hosted a Google+ Hangout about the potential of nature's microbiomes and how we can tap into it.

Gut microbes

Why It's Time to Map the Microbiome

Learn more about the Unified Microbiome Project in an earlier discussion with Janet Jansson, Rob Knight and Jeff Miller in which they discussed how such a national research effort could impact human health and the environment. Or watch the video of the Hangout.


About the participants (left to right):

  • JANET JANSSON – is chief scientist of biology in the Earth and Biological Sciences Directorate at Pacific Northwest National Laboratory (PNNL) and sector lead for PNNL research in the Department of Energy's Biological Systems Science Division. She coordinates two of PNNL’s biology programs: the Microbiomes in Transition (MinT) initiative to study how climate and environmental changes impact natural and human microbiomes and the DOE Foundational Scientific Focus Area, Principles of Microbial Community Design.
  • ROB KNIGHT – is the founder of the American Gut Project, an open-access project to survey the digestive system’s microbiome and its effect human health and development. He holds appointments at the University of California, San Diego School of Medicine and Department of Computer Science and Engineering, where he develops bioinformatics systems to classify and interpret large sets of biological data.
  • JEFF F. MILLER – is director of the California NanoSystems Institute, a multidisciplinary research organization, and the corresponding author of the consortium’s Science paper. Based at University of California, Los Angeles, Miller holds the Fred Kavli Chair in NanoSystems Sciences and is a professor of Microbiology, Immunology & Molecular Genetics.
  • ALAN BROWN (moderator) – is a freelance journalist and writer who specializes in science, engineering, and technology. He has been covering nanoscience and nanotechnology for more than 25 years.

Below is a modified transcript of the discussion. Edits and changes have been made by the participants to clarify spoken comments recorded during the live webcast.

The Kavli Foundation: So, let's start with a question. There's been a Cambrian Explosion in microbiome research. Ten years ago, microbiomes were hardly on the map. Last year, 25,000 papers contained the term. Why is this happening now? Is it just because we can read microbial DNA, or are other technologies making this possible?

Jeff MillerJeff Miller, Fred Kavli Chair in NanoSystems Sciences at the University of California, Los Angeles. He studies the molecular mechanisms of bacterial diseases and the evolution of bacterial diversity. (Credit: UCLA)

Jeff Miller: There are a lot of factors that came together to cause this explosion of interest. One, certainly, is the ability to rapidly sequence DNA. And over the past 10 years or so, we've seen a progression of technologies that allow us to characterize microbial communities with increasing resolution and sophistication. But we've also encountered many bottlenecks along the way. And interpreting this massive amount of sequenced data is one of those bottlenecks.

Rob Knight: I agree. I think it's really the combination of the DNA sequencing tools getting much cheaper, and the computational tools, including the toolkits that we developed, that make the information much more accessible to a broad community of users. I think what we will see in the future are tools that will go beyond taking inventories of species or inventories of genes and instead provide much more insight about how these species and genes function. But that is going to require a whole lot of additional development of both the software and the knowledge base to use that software.

TKF: Janet, do you have any additional thoughts on that?

Janet Jansson: Yes. With DNA sequencing we get information about the composition of microbiomes, but it's also interesting to know what those microbes are doing. For example, if we could understand their protein or metabolite composition, we could get a better understanding of what they're doing in different kinds of habitats and inside our bodies. There are a lot of developments in these areas, but those tools are still lagging behind the sequencing technologies.

TKF: So, do we need a major program, a Unified Microbiome Initiative, to develop these capabilities? Couldn't we build on existing technologies or do we need to invent radically new types of science?

Miller: The likely answer is, "both." There's certainly a lot of room for incremental advances leading to better sequencing technology and the like. But we also need some quantum leaps at the same time.

The field has progressed rapidly. But we've reached a plateau that has to do with the limitations of the current technologies. We need to be able to see microbial communities where they live, in real time. We want to know what are they doing. What genes are they expressing? What proteins are they making? What metabolites are they synthesizing? How are they responding to each other and their environments?

Then we need to be able to take all this data and interpret it in a way that allows us to ask questions and formulate new hypotheses that we can test and falsify or prove correct.

These are really tall orders. They're going to require not only new technologies, but also the input of collaborators in engineering, physics, and chemistry, as well as the life sciences, environmental sciences, computer sciences, and more.

TKF: I'm curious about the computer science side of it. Rob, you have a joint appointment UC San Diego's medical school and computer science department. Is it such a tall order? I mean, we have big data. Are we going to need something more?

Rob KnightRob Knight is founder of the American Gut Project, a crowdsourced research program that is analyzing thousands of human gut microbiomes. He holds joint appointments at the University of California, San Diego, School of Medicine and Department of Computer Science and Engineering.(Credit: UC San Diego)

Knight: Well, the issue is that big data and magic are not quite the same thing. There's a lot of advances that need to happen on the algorithm side. In general, machine learning and generic algorithms will give you a good, but not ideal, answer to a particular scientific question. And the more information you can put in at the beginning to tailor those algorithms to your specific problem, the better you'll do.

The other thing is that although we're producing a tremendous amount of data, we're still limited by the amount of data—it’s still not enough—and by our ability to interpret it. The problem a lot of people are facing right now is that they have collected so much microbial community information. They have over a thousand species that they don't understand. They are listing a million genes they don't understand. Then they are going onto measuring other types of molecules using metatranscriptomics or metaproteomics or metabolomics where, again, they create very large inventories that they also don't understand.

But even with all that data, we're still limited by the number of samples, and by our ability to annotate and understand those entities. There's a huge role for both existing algorithms that can be applied more effectively as we get more data, and for fundamentally new algorithms as well as new ways of computing that radically change how we think about computation itself.

TKF: Part of the challenge is that we need a better way to get closer to the inhabitants of metaphorical city I mentioned earlier. It is as if we are looking at that city from space and trying to figure out people's roles when we cannot even see these individuals, isn't it?

Knight: It's a little worse than that. You're flying out there in your UFO, and you just take a big chunk of that city, grind it up, look at all the DNA and chemicals, and try to make sense of it. That can be an effective or ineffective way to understand the city. You will get an understanding of some of the chemical processes that are going on, and some of the genes that are expressed. But you're not going to learn a lot about the sociology or how those organisms communicate.

Jansson: Yes, and another way to tackle that problem is to use simpler model communities. That way, if we don't have the instruments and data tools to deal with these highly complex communities, at least have a model community that will let us study specific interactions.   

“We've already radically reengineered, through agriculture, both soil and human microbiomes over most of the planet. We're brought them into states that have no precedent in nature.”—Rob Knight

TKF: In other words, it's easier to study something much simpler?

DiatomsResearchers at PNNL are developing new tools to study marine microorganisms such as these. (Credit: NOAA)

Jansson: Yes, at least for now. Full communities are some of the most diverse types of habitats for microorganisms on earth. We go so much data, that we're not limited by the amount of data we produce, but by our ability to process the data. Even with supercomputers, it can take weeks, if not months, to just run all of that data through our computers.

Knight: With all due respect, I think we're still data limited because we don't have enough samples.

So, it's as if we had, say, five photos, and we're taking them at higher and higher resolution. That generates a lot of data, but not enough to create a movie. What we really need is, say, 100,000 frames. And no matter how much more information we get out of the smaller number of frames you have, we'll never be able to put that movie together.

So, that's a lot of what we're facing. Right now, it's so expensive to process each sample, it's really difficult to get enough samples. This is really why we need to be able to read out microbes much, much faster, much, much cheaper. And we also need to use higher and higher resolution techniques, to get that full movie of how the interactions are taking place.

Jansson: I agree we need more samples. But even then, it's very difficult to process the information from one sample.

Miller: Right. In fact, we know the functions of only about half of the genes that we detect in these communities. And of the half we think we know, the amount of mis-annotation and improper context annotation, are also significant. So we're trying to put a puzzle together with only some of the pieces. And if you look at small molecules, this situation is even worse. About two percent of the metabolites that are found in the typical microbial community map to known structures. And only a fraction of those two percent are on known biochemical pathways. So we need more information.

"The concept of a species has to be reconsidered when you're talking about microbes, because they're not only diverse, but constantly exchanging genetic infrormation."—Jeff Miller

TKF: Those metabolites are involved in bacterial digestion. Are they how bacteria communicate with one another?

Miller: Yes, that how they communicate, and how they acquire energy. They are the waste products they release, and the small molecules they use to compete with other microbes and interact with their environments. And many other things that have yet to be discovered. These small molecules are the language of microbial communities.

TKF: Getting a handle on all this sounds like an imposing research project. But suppose you had these tools today. What is it that you'd like to study? Jeff, you study the evolution of bacteria that cause disease. What would you do with those tools?

Miller: Boy, great question. I think one area that is prime for progress -- and some progress has been made already -- is the idea of taking a community that may be somewhat robust but not really optimal for its environment or host and engineering it so that it has more beneficial properties and fewer non-beneficial properties.

Doing that really requires an understanding of the ecological principals that govern the community's composition, robustness, response to changes, etcetera. So, being able to reprogram microbial communities is really one of our ultimate goals.

There are various steps along that pathway that one can imagine. But we're just at the very early stages of being able to do that. So if I were to choose one thing to study, it would be to understand how microbial communities are constructed well enough to enable predictive reliable, reengineering of those communities in order to optimize their functions.

TKF: Very interesting. Janet, I know you collaborate on human microbiome work. But you've also developed a reputation for investigating how environmental changes affect microbiomes in the Alaskan permafrost and on the Gulf of Mexico. What types of things have you learned and what would new tools tell you that you don't already know?

Janet JanssonJanet Jansson is chief scientist of biology in the Earth and Biological Sciences Directorate at Pacific Northwest National Laboratory, where she leads the effort to study microbial communities in soil, sediment and the human body. (Credit: Pacific Northwest National Laboratory)

Jansson:  For environmental studies, we want to understand how events, such as the Deepwater Horizon oil spill on the Gulf or the thawing of permafrost due to global warming in Alaska, is impacting microbes and the processes that they carry out in those systems.

With the Gulf oil spill, we had organisms that were enriched during the spill, and that were able to degrade oil. So that was interesting, from that perspective.

In the permafrost, we have a huge reserve of carbon that is currently trapped in that environment. So what happens to that carbon as the permafrost thaws and the microorganisms start to become active and degrade the carbon? Are they going to release a lot more carbon dioxide to the atmosphere and make the global warming process worse? At a very fundamental level, we need to understand what those microorganisms are doing.

TKF: Very good. I'd like to move to some listener questions. You know, microbiomes are suddenly in the news, and several listeners want to know about products that promise to improve our health and cure certain conditions by altering our microbiomes. Rob, you've been studying the American gut for a while now. Do we know enough about microbiomes for anybody to make that kind of a claim?

Knight: Yes, but so far, that's limited to just a very small number of people. For example, there was a really nice paper in Cell by Eran Segal and Eran Elinav of Israel's Weizmann Institute of Science. It showed that based on your microbiome, you can predict what foods will have good or bad impacts on your blood sugar. The drawback, so far, is that they can only do that in the Israeli population, where the food item inventory is somewhat different from what you would see in the United States, for example. But that technology is on the horizon and improving very rapidly.

As far as probiotics go, there's not a lot of evidence that probiotics improve general health in humans, though there is some intriguing data in mice. On the other hand, there's a fair number of probiotics that have been clinically studied in well-conducted randomized controlled trials. For a number of conditions, like, irritable bowel syndrome, post-antibiotic diarrhea, and so forth, there are particular probiotics on the market that have been clinically validated.

However, it's kind of like drugs, where certain probiotics are good for particular conditions, but not something that you should take generally. And in the same way that you would expect for drugs, most people don't need to take most probiotics most of the time, or at least not the ones that have been studied so far. So, I think it's fair to say that public enthusiasm is greatly outstripping the actual evidence. But there is some evidence underlying that enthusiasm.

TKF: Jeff, what about the future? Are we going to be able to cure diseases? Will I be able to speed up my microbiome's metabolism so I can eat ice cream and never gain an ounce?

Miller: When you look at the probiotics that are out there, they date way back. They have their origins in food production, fermentation, cheese making, and other processes. So the question is, do they have a health benefit or not? And the results are often equivocal.

But that's very different than looking at what we know now, and asking, okay, how would you engineer or reengineer this system? Would a small consortia of bacteria be a good way to decrease fatty tissue and increase muscle mass with diet? So, as Rob said, we haven't yet gotten to the point where we have applied our modern understanding of microbiomes to probiotics now in the marketplace. But the potential for doing that is definitely there.

So, to answer your question, it could cure infectious diseases. A great example is Clostridium difficile-induced diarrhea, which is caused by antibiotics. The best cure that we know is fecal microbiome transplantation from a healthy donor. It is about 90 percent effective, so we know it can work. It's very crude, and so the question now is how to make it better through more refined science, rather than hit-and-miss empirical testing.

Knight: It's important to remember that this is not just for the future. There are people walking around, alive now, who would be dead had they not received fecal microbiome transplants. This is really a current technology that works and is being clinically applied now. And what we need to do is to refine it. But it's not something that's in the future, it's something that's here today.

“So what happens to that carbon as the permafrost thaws and the microorganisms start to become active and degrade the carbon? Are they going to release a lot more carbon dioxide to the atmosphere and make the global warming process worse?”—Janet Jansson

TKF: This opens up some very interesting questions. One of the things we've discovered about the human microbiome is that it influences all kinds of things, from brain development and obesity to behavior. These are the very things that define who we are. Now we're talking about possibly synthesizing artificial microbiomes. This raises some ethical issues, doesn't it?

Miller: Definitely. Ethics is a huge, huge area. "Do no harm" is the first principal, whether we're talking about permafrost, agriculture or the human gastrointestinal tract. And so, the requirements for reengineering microbiota to use as a drug have got to be stringent and carefully controlled. And safety, obviously, is going to be the first issue.

But it's complicated, because these are dynamic systems. And the question is, how long will any changes last? What else would change the result of making these perturbations, etcetera? So we need to understand a great deal more before we try to engineer and manipulate at a large scale.

TKF: Janet, you study ecology. Could you imagine a large scale ecological intervention using microbiomes?

BeckmanUCICarbon in some types of ancient permafrost is digested by greenhouse gas-producing microbes.

Jansson: Before I address that, I just want to go back to our earlier discussion about probiotics. In addition to changing our microbiome, we can also influence it through the food we put into it. This is also a strategy that is sometimes successful, though not very well understood. Instead of a probiotic, it’s called a prebiotic. For example, you can eat what is called a resistant carbohydrate or starch, which is not easy to digest. So it makes it to your intestine relatively intact. This allows the microorganisms in your gut to consume and ferment it, and that's beneficial for the colonic health.

As for actually manipulating an ecosystem on a large scale, this is, of course, difficult. There have been people who have talked about fertilizing the oceans by adding iron, to buffer or mitigate the impact of increasing CO2 concentrations. But when it comes to permafrost, how to prevent the degradation of the carbon that is trapped there? That is difficult. But by gaining knowledge about the types of organisms that are there and the ones that become active when the permafrost does begin to thaw, we can at least predict the implications of those changes.

Knight: Just to build on what Janet said, it's important to remember that we've already radically reengineered, through agriculture, both soil and human microbiomes over most of the planet. We're brought them into states that have no precedent in nature.

The issue is that we didn't understand at all what we were doing or what our impacts on those microbiomes were. So, it's not that we can't change them. We are already changing them. And have already changed them. The question's more, "Can we change them in a more nuanced and directed way, where we have a better understanding of the ways that we can change them, at the microbiome level as opposed to the industrial or occupational level?"

TKF: We've talked about microbiomes impacting out development and behavior. These are the things that determine our personality. For a long time, researchers thought that our genetic makeup determined these things. Do we understand the interaction between microbiomes and genome? Janet, you're shaking your head, so why don't you start.

Jansson: I can tell you that this is a real hot area of research right now. My group and several other groups are trying to establish the link between the host's genome and the microbiome. I can say that preliminary evidence – there have been a few publications mainly looking at mouse models – suggest that there is a link. Rob's taken a more historical perspective, looking at different types of human populations and the impact of ancestral lifestyles on microbiomes. Rob, maybe you want to comment on that?

Knight: Yes. We know that both in mice and in humans, lifestyle behaviors, like diet and hygiene especially, have had a much larger impact than host genetics. This is true, even though host genetics still has a highly statistically significant impact on particular features of the microbiome, including, interestingly, features that are associated with obesity in humans.

Miller: To add one thing to what Rob said, we've coevolved with our microbial communities since long before we became Homo sapiens. We have only about a dozen genes in our genome to digest complex carbohydrates. The microbiota in our gastrointestinal tract brings hundreds of genes that do that for us. And so, when we eat a healthy high fiber diet, what we're really doing is relying on these microbial consortia to digest that food for us, so that we can take some of the products and use them for energy and other purposes.

"A decade ago we had hints that the microbiome was linked to health. But no one predicted, at all, that it would have a key role in behavior, especially in mammals."—Rob Knight

TKF: So, as one listener asks, maybe it's not such a great idea to use bactericides to kill microbes on every surface in our homes?

Miller: Not a good idea for a lot of reasons. Rob, you shook your head, so I'll let you start.

Knight: Well, it's bad for so many reasons. Both in terms of increasing antimicrobial-resistant bacteria, because the bacteria that survive your attempts to kill them can then spread those resistant genes to other bacteria that infect us directly. And also because there's evidence, increasingly, that keeping your house too clean increases the risk of autoimmune diseases, especially in children.

TKF: We're drawing to the end of our discussion, so I want to ask you a final question. You know, our understanding of the microbiome has changed dramatically over the past 10 or 15 years. Tell me, what has surprised you most about what you have discovered? Janet, why don't we start with you?

Jansson: I think the thing that has surprised me the most is the importance of the microbiome with respect to our health, in so many different ways. This was something that was not known at all just a decade ago. And so that's what I'll say.

TKF: Okay. Rob?

Knight: Links between the microbiome and behavior. A decade ago we had hints that the microbiome was linked to health. But no one predicted, at all, that it would have a key role in behavior, especially in mammals.

TKF: And Jeff?

Miller: Diversity. Microbes – whether you're studying pathogens, beneficial microbes, or microbes in any context – are enormously diverse. The concept of a species has to be reconsidered when you're talking about microbes, because they're not only diverse, but constantly exchanging genetic information. They are truly a constantly moving target, and the extent of their functional diversity is mind-boggling.

TKF: Excellent. This is certainly an exciting time for microbial research. And I didn't even get to ask the best question, which is, “How does the microbiome in our gut determine our behavior?”

Knight: We don't know how it happens, and that's why we need a Unified Microbiome Initiative.