Swabbing for Science

    Escherichia coli (E. coli) bacteria (blue) in the small intestine of a child (yellow). Credit: Stephanie Schuller

    Escherichia coli (E. coli) bacteria (blue) in the small intestine of a child (yellow).
    Credit: Stephanie Schuller

    Scientific breakthroughs commonly change how we see the world, but uBiome, a Bay Area startup, wants to change how those breakthroughs happen.

    uBiome is fittingly pronounced “you-biome,” and for a small fee, they’ll sequence your microbiome: the collection of microbes that nestle in the nooks and crannies of your body. When considering the microbiome, it helps to relinquish a human-centric view. Instead, think of the human body as a mobile ecosystem, with thousands of species—of which there is only a single Homo sapiens—working together to stay alive. The interplay between our health and the microbiome has recently captured the imagination of popular science. Articles in the New York Times, the Economist, and the New Yorker refer to the microbiome as a human organ, citing estimations that there are 10 times more microbial cells associated with our bodies than human cells. Like any respectable organ, the microbiome weighs several ounces, has many cell types, and communicates with other organs.

    And like all organs, the microbiome is essential to our health. Many plant carbohydrates are indigestible by human enzymes but can be digested by bacteria that live in our gut. We rely on bacteria to process a tenth of our consumed calories. A healthy microbiome is also vital to our development: in infancy, we need exposure to the correct types of microbes to train our immune systems to recognize invading pathogens. The microbiome may even alter brain chemistry, which would allow it to influence mood and emotion (see “Manipulative Microbes,” BSR  Fall 2012).

    Although personalized analysis of human genomes is commercially available, each microbe has its own genome, which means that the human genome is just one out of trillions involved in a person’s health. This is where uBiome can help. They will harness three joint enthusiasms to launch their company: crowdfunding, which relies on the willingness of strangers to give money for a cause that appeals to them; citizen science, which engages a layperson’s curiosity about the world; and personalized medicine, which attempts to tailor medical treatment at a genomic level—even if the genomes in question are those of microbial stowaways.

    The birth of a company

    uBiome’s inception is rooted in a natural curiosity about our own health. Over lunch one day in 2012, William Ludington and Zachary Apte discussed an urge to examine their own microbiomes. As graduate students at UCSF, both were surprised at how difficult it would be to obtain that information as a member of the public.  Ludington explained, “We really wanted to sequence ourselves, but there wasn’t a cheap place to do it that would give us meaningful, quick answers.” With the expertise of cofounder Jessica Richman, a graduate student at Oxford University and self-described serial entrepreneur, they founded uBiome in the California Institute for Quantitative Biosciences (QB3) Garage, an incubator that provides support and space to bud research from UC labs into new companies (“A lab space of one’s own,” BSR Fall 2011). All three founders lived in the Bay Area, but Apte had met Richman abroad though Startup Chile, an entrepreneurship program established by the Chilean government. Ludington is now a Bowes Fellow in the Department of Molecular and Cellular Biology at UC Berkeley, Apte is pursuing his interest in startups with the QB3 incubator, and Richman is continuing her studies in England.

    uBiome’s business model is simple. After pledging a certain level of financial support, each uBiome customer will receive a collection kit to swab a sample of the bacterial population on his or her body. Like nail polish bottles, the tube caps have an attached sample brush (which resembles a Q-tip), and are color-coded for each sampling site: green for the gastrointestinal tract, blue for noses, red for mouths, yellow for behind the ears, and purple for genitals. After swabbing and completing a survey, customers send these kits back to uBiome, where they’ll sequence DNA from the cells collected in the swab samples.

    Instead of sequencing entire genomes, uBiome will keep costs down by sequencing only one gene: the 16S ribosomal RNA gene, which is commonly used for tracing relationships between bacterial species. Identifying which sequences are present in a sample will allow uBiome to infer, at the genus level, which bacteria are present in the population. This technique has the added advantage of singling out bacterial DNA from that of human or other microbes present in the sample (whose genomes lack this gene). However, this method doesn’t allow for sorting bacteria at the species or strain level. These distinctions are small, but can have significant health impacts. Although many strains of E. coli live in mammalian intestines, only a few notorious strains cause illness, despite sharing most of their DNA with innocuous E. coli.

    After sequencing, uBiome will send customers a link to their website, where they can view a profile listing the types and relative quantities of bacteria living on their sampling site. Online, customers can compare profiles and survey answers to those of other uBiome customers, and read published studies that may be relevant to their profile. Customers can opt to include their data (anonymously) in uBiome’s database, which the company plans to make publicly available. Dr. Liam Holt, a Bowes Fellow at UC Berkeley and a scientific advisor to uBiome, admitted that uBiome’s goal was “not that lofty, in a sense. Where there’s really very little information about microbiomes in the population, we can at least get some.” Their database will form the core of an online community, which they hope will demonstrate that citizen science can guide meaningful research.

    The science of microbiomes

    Design: Asako Miyakawa Data: Nature, NIH

    Design: Asako Miyakawa
    Data: Nature, NIH

    When presented with the idea of a microbe, most people imagine tiny invaders that wreak mischief on their host. This is because many medical breakthroughs in the 20th century have been in response to bacterial infections, which antibiotics are very effective at eliminating. But chronic diseases have proven much harder to treat, and bacteria have been implicated in illnesses like diabetes, autism, and a wide range of autoimmune disorders. This kind of research isn’t as simple as sorting bacteria into categories like “beneficial” or “pathogenic.” Having the right kinds of bacteria, in the right amounts, helps to keep potentially pathogenic but otherwise innocuous bacteria in line. Holt described this set of interactions as an ecosystem like any other, with “each individual microbe in a microbiome dependent on a network. When you have a disease, that’s often because the network properties are messed up.” Population dynamics within a microbiome can tip the balance towards health or illness, and this presents a problem that can’t be cured with a simple antibiotic.

    The recent completion of the Human Microbiome Project, an NIH-funded effort to sequence microbiomes in both the healthy and diseased, has highlighted the dizzying variety of bacteria that make up our microbiome. An original goal of the Human Microbiome Project was to define a “core microbiome”—a subset of bacterial species that most people have in common—but now researchers aren’t sure whether it even exists. In the gut, microbiome profiles differ widely from individual to individual, and someone’s profile can change drastically from day to day. These differences between and within individuals make potential therapies tough to untangle.

    Each customer will receive a collection kit for swabbing a sample of bacteria from one of five sites on their body (the tubes are color coded for each site): green for guts, blue for noses, red for mouths, yellow for behind the ears, and purple for genitals. Photos: uBiome Design: Asako Miyakawa

    Each customer will receive a collection kit for swabbing a sample of bacteria from one of five sites on their body (the tubes are color coded for each site): green for guts, blue for noses, red for mouths, yellow for behind the ears, and purple for genitals.
    Photos: uBiome
    Design: Asako Miyakawa

    But producers of probiotic supplements aren’t afraid to try. Probiotics are live microorganisms that may have a beneficial effect on health, and come in forms that range from daily pills to cups of yogurt or even bottles of kombucha. Most probiotics consist only of a few bacterial species, and it isn’t clear how the introduction of a few species affects the complex ecosystem of the microbiome. Very few probiotics have passed preliminary clinical trials, which makes it difficult to evaluate their claims. To date, the European Food Safety Authority has rejected more than 260 claims on probiotic supplements due to inconclusive proof from trials.

    A more promising therapy for microbial imbalances in the gut is also much less appetizing: fecal transplants, in which a doctor transfers feces from a healthy donor to the patient with a suppository. In a 2013 study, fecal transplants (dryly referred to in the paper as “a duodenal infusion of donor feces”) were 94 percent effective at treating patients suffering from recurrent diarrhea caused by Clostridium difficile infection. The control groups tested two different antibiotic treatments, and each was only a third as effective.  The efficacy of fecal transplants wasn’t lost on those in the control groups—almost half of them went on to have “off-protocol” transplants after the study was completed. The researchers noted an explosion of diversity in gut bacteria after transplantation, which underscores the importance of diversity in microbiome health, not just the presence or absence of a few bacterial species.

    The microbiome’s malleability makes this research seem encouraging, especially when compared to using human gene therapy to treat disease. “It’s difficult to change our genomes, so this may be the next best thing,” said Ludington. Or as Holt put it, “You can eat a yogurt and change [your microbiome], which is easier than injecting engineered stem cells into your body.” The future of medicine may rely on specific therapies that combine probiotic supplements with targeted antibiotics, which would work like assassins instead of bombs to eliminate illness.

    Self-knowledge through numbers

    uBiome’s mission, according to Ludington, is to “develop the technology and establish a pipeline so that anyone can do microbiome research.” For the first time, uBiome will allow those who have made lifestyle changes (perhaps a new diet, exercise regimen, or facial cleanser) to follow microbial consequences over time by submitting new samples to uBiome. They envision their customers performing the types of casual experiments that are suggested during happy hour but rarely make it to a lab. For uBiome customers, gratification will come in the form of concrete data. Ludington explained that “uBiome provides a way to evaluate changes critically.” When asked about fad diets or expensive cosmetics, he commented, “We all do these voodoo things. If we could check whether they actually worked, we could avoid wasting our time.”

    uBiome may appeal most to those who are frustrated with their current state of health care. “Customers with gastrointestinal tract or sinus problems can ask their own questions about the role their microbiome plays in these problems. This could eventually lead to more understanding and personal control over their health,” said Ludington. The medical establishment lacks concrete solutions for many communities brought together by disease, or sometimes just common symptoms. Those centered around autism, bowel disorders, and allergies all have vocal online presences. To counter their frustrations, some members of these communities have turned to the comfort provided by the so-called “quantified self” movement, whose motto is “self-knowledge through numbers.”

    By tracking personal metrics (weight gain, exercise, and sleep patterns are three of the most common), quantified-self enthusiasts wish to determine the effectiveness of unusual therapies. When patients choose a medically unconventional therapy, their only option is to plunge into treatment without the protection of clinical studies or an understanding of biological mechanism. A smartphone can help patients keep track of sleep patterns or caloric intake, but some metrics like blood cell counts or lipid panels require sophisticated technologies. uBiome expands the scope of personal data by providing a relatively inexpensive method to track changes in a microbiome.

    Richman portrayed their community as a haven for “people who are passionate about a condition to find others online and share their experiences” in the context of their microbiome profiles. Even if microbiome data can’t identify an illness or prescribe a cure, uBiome empowers individuals to take matters into their own hands. At the least, these types of personal experiments will grant uBiome customers agency over their own health—something the medical establishment can’t always provide.

    Based on the relative proportions of bacterial species in each sample, uBiome can cluster people into similar enterotypes (subgroups of bacterial ecosystems). This network represents several enterotypes for each of the possible sampling sites. The larger circles are the more common enterotypes, while smaller circles are more rare. The lines between them represent the similarities between each enterotype.

    Based on the relative proportions of bacterial species in each sample, uBiome can cluster people into similar enterotypes (subgroups of bacterial ecosystems). This network represents several enterotypes for each of the possible sampling sites. The larger circles are the more common enterotypes, while smaller circles are more rare. The lines between them represent the similarities between each enterotype.
    Credit: uBiome

    Turning anecdotes into data

    Although uBiome allows patients to obtain their data without a physician’s guidance, their website is straightforward about the limits of their service. “We are not [medical] doctors and cannot diagnose medical conditions. uBiome is not a diagnostic test.” They encourage patients to take the information provided by uBiome to their doctor, to “discuss the results of the test and determine the best way to proceed.” This advisory is a little premature—since microbiome research is still in its infancy, there are no medical guidelines that instruct doctors on how to use the data. Ludington admitted that the current therapeutic scope of his company is nonexistent. “There aren’t a lot of cures for microbiome ailments—at this point, all uBiome can do is tell you what your microbiome looks similar to.”

    Taken singly, an individual’s data lack external controls (comparisons to other people) and statistical power (with sampling from just one person). The real power of microbiome data lies within the community uBiome hopes to create. When a single experiment is repeated in many individuals—for example, avoiding gluten—the larger sample size can weed out random noise and perhaps identify lifestyle changes that could affect both microbiome composition and health. Richman described this as the “point at which a collection of anecdotes turns into data.” She believes that uBiome, and citizen science in general, can “help to legitimize the anecdote through rigorous analysis and combining it with others’.”

    “That’s the name of the game: statistical power,” declared Holt. “What can we discover with 2,000 people that we can’t know for sure with 200?” Based on the prevalence of these diseases in an American population, uBiome will probably receive samples from those with diabetes, arthritis, and asthma, and may even have some from those with breast cancer, dementia, or Crohn’s disease. If they could get tens of thousands of swabbers, uBiome would likely be able to analyze profiles of those with multiple sclerosis, autism, and a host of cancers. “It sucks that we have to charge money for the kits,” Ludington joked, half-seriously. “If we could make it free, we’d get a lot more people involved,” which would increase the chance of uncovering correlations that could lead to therapies. This may become a challenge for uBiome as it navigates a path between profit-driven company and science-driven research organization.

    A novel pathway for funding

    In academia, projects are funded on their ability to predict a clear scientific innovation. Grant reviewers appreciate clear hypotheses. But in practice, science heads in unplanned directions that can have profound impacts, such as the Nobel-winning discoveries of RNAi (see “Small Interference, Big Impact?” BSR Spring 2012) or green fluorescent protein (GFP). uBiome is taking the path of no hypothesis, which makes it difficult to get funded by traditional methods. “It’s all discovery-based,” explained Holt. “There’s a zero chance you’d get any money from the NIH to do this.” He’s probably right: none of the founders have a publication record in this field, they aren’t testing explicit hypotheses, and research funding gets more competitive every year.

    After sequencing the bacteria in your sample, uBiome will provide a profile that compares it to others’. By examining the similarities (the colors of each box) between the relative amounts of each bacterial genus, they can determine relatedness between samples, shown in the tree to the left of the matrix. Credit: uBiome

    After sequencing the bacteria in your sample, uBiome will provide a profile that compares it to others’. By examining the similarities (the colors of each box) between the relative amounts of each bacterial genus, they can determine relatedness between samples, shown in the tree to the left of the matrix.
    Credit: uBiome

    This is why uBiome has turned to a new method of raising money: crowdfunding, which has been popularized over the last few years by websites like Kickstarter and Indiegogo. It’s a relatively simple way for small ventures to get the initial cash needed for their projects. Anyone can donate a few dollars to a project in exchange for the thrill of being an early adopter, a sense of community with other donors, and often a perk, like access to the finished product.

    uBiome decided to launch its funding campaign on Indiegogo, a website that doesn’t curate projects or comment on the trustworthiness of the creators—it allows the crowd to decide on the worth and dependability of a project. Since crowdfunded projects aren’t beholden to investors, they have more freedom to experiment with risk, which also grants them more freedom to fail. Crowdfunding investors don’t own equity in the company and have no control over the direction of the project. On Indiegogo, contributors to a project assume the entire risk of investing—the website makes it clear that they do not grant refunds for failed campaigns.

    For its first round of fundraising, which ended in February 2013, uBiome set its goal at $100,000. This would be a modest sum for venture capital, but a lofty target for crowdfunded projects, which typically ask for a few thousand dollars. To meet their goal, they reached out to a variety of online communities: geeks who comprise the do-it-yourself community (with mentions in Wired, Scientific American, NPR, and the Huffington Post) and those brought together by an interest in health or a common disease (by posting on forums for “paleo” dieters, alternative health treatments, or gastrointestinal ailments like celiac or Crohn’s disease). They also offered many tiers of commitment, from $5 to simply show support up to $1,337 to sample five sites at five different time-points.

    uBiome’s marketing efforts helped them raise $351,193, wildly exceeding their funding goal. This figure, their website claims, makes uBiome “the largest successful crowdfunding campaign for citizen science to date.” Holt pointed out that uBiome now has more money than a typical R01 grant, the most common NIH research grant, and that uBiome raised that amount in just three months, while an NIH grant cycle can take up to a year. Ludington saw it all as part of their company’s mission. “We’re not just doing science, but citizen science—we directly involve the public in doing their own scientific research, and we can reach out to them immediately.”

    Science by the people, for the people

    The Human Microbiome Project sequenced 242 individuals, but uBiome plans to sequence thousands to build the largest set of microbiomes in the world. The first wave of crowdfunded sales will send kits to over 1,700 people. But uBiome doesn’t want public involvement to end with monetary or sample contributions, because they’re also experimenting with crowdsourcing their research. As Ludington described it, “We want the public to drive.”

    Citizen science has benefited from the connectivity of the Internet, but it’s hardly a new concept. Since the 1900s, the Audubon Society has relied on bird counts submitted by amateur bird-watchers to compile population data for many North American species. Some citizen science efforts have already passed the standard of academic rigor by making it to publication. David Baker, a scientist at the University of Washington, designed the puzzle game FoldIt to crowdsource protein structure predictions. In 2011, FoldIt players only took 10 days to solve the structure of a protein involved in simian immunodeficiency virus, a protein whose structure had remained unsolved for 15 years.  This collaboration resulted in a Nature paper for the FoldIt players and designers.

    Richman sees a flaw in the way scientific research is currently conducted, and believes citizen science presents a solution. With the continued specialization of scientists that began in the 20th century, she saw the gap widening “between how many people consume science and how many people produce science.” Clear communication from scientists to the public is one way to bridge this disparity, but Richman feels that “in a more interactive world, education comes as much from active involvement as it does from being lectured to by experts.” She would like for uBiome’s community to surpass previous citizen science efforts (where participants can contribute a sample or play a game for analysis) to demonstrate that amateurs can successfully guide research. “There are so many intractable problems in the world for which science is the best way of addressing them. To add all that extra human energy into solving those problems is just a great opportunity.”

    Of course, the benefits of community-based research can also become pitfalls. Citizen scientists, by definition, aren’t highly trained, and may introduce unexpected biases or flawed data. Crowdsourced ideas rely on popularity to distinguish worthiness, but in science, a correct answer can be unintuitive, and therefore unpopular. And although most citizen science efforts have managed to avoid the vitriol found in some corners of the Internet, a crowd can easily become a mob in which private interests overrule the public good.

    uBiome will utilize the popularity of the do-it-yourself spirit to create an online community where anyone can suggest a hypothesis or run a new analysis on their data. With such a large dataset, independent researchers will be able to develop or train analyses with more specificity. Researchers can choose to make their algorithms and visualizations available on uBiome’s site, which would grow into a repository of user-friendly tools. Since the community doesn’t exist yet, no one can say what form it will take, but Ludington envisioned a site with both an informational and social media aspect, “kind of like Reddit, where people can upvote or downvote questions to see what’s most interesting to the community.”

    uBiome customers have already posed questions that its founders hadn’t anticipated. They’ve been contacted by families with histories of disease who want to conduct generational studies in their family tree. Richman mentioned a transgendered woman who asked her about the effects of the sexual transition on microflora, which has only been examined in a few studies. Holt was aware of individuals who are tracking their microbiome over their lifetimes, keeping corresponding diaries of lifestyle habits and microbiome profiles. To facilitate these sorts of longitudinal studies, he thinks it would be interesting “to develop an even more cost-effective way to generate these profiles, perhaps by focusing on a few key bacterial species.” Ludington is on the same page. His son was born in December 2012, and may become the most thoroughly sampled individual alive. “I just took poop sample number 54 this morning,” Ludington commented in January, and he had no plans to stop.

    uBiome founders Jessica Richman, Zachary Apte, and William Ludington clearly love science. They hope that uBiome, by involving the public with funding and directing microbiome research, can infect others with their enthusiasm. Credit: uBiome

    uBiome founders Jessica Richman, Zachary Apte, and William Ludington clearly love science. They hope that uBiome, by involving the public with funding and directing microbiome research, can infect others with their enthusiasm.
    Credit: uBiome

    The wisdom of a crowd

    In its simplest form, uBiome is a new company that will sequence your microbiome. But in doing so, it will also provide a forum that folds the public into the traditionally mysterious scientific process. Richman observed that some attempts to bridge the divide between academia and the public can be a little patronizing. “It’s not that we don’t need experts,” she explained, “but there needs to be more of a continuum between a layperson and a professional scientist.” It’s the lack of a hypothesis—the very thing that makes this sort of research difficult to fund—that could even generate more public enthusiasm. uBiome allows citizen scientists the freedom to imagine an infinite number of discoveries from their research, which makes the most unconventional (and scary) thing about uBiome the most exciting.

    “This kind of study is very immediate,” said Holt. “It’s interesting, this different style. We start running first, and then see what happens.” If microbiome research doesn’t advance quickly enough, the information uBiome provides is unlikely to be helpful, which makes their hopes for an engaged community seem far-fetched. Even if the research—either academic or amateur—looks promising, population dynamics can be very difficult to interpret, especially in diagnosing nuanced medical conditions. But uBiome has situated itself to take advantage of the “energy, intelligence, and passion” that Richman believes lies within many potential citizen scientists. Ludington was optimistic about what he calls the “wisdom of a crowd,” which isn’t guided by any plan at all. “If the research heads off a cliff, it’s inconsequential. But if it goes somewhere, that’s awesome – that’s science driven completely by the public.”

    Leave a Reply