He cites two reasons for his optimism: genetic diversity and treatability. One person’s genome differs only 0.1% from another’s; while their gut genomes may differ by 50%. “Since there’s so much variability, there is a much greater chance we’ll be able to associate differences with disease,” he says, adding that it should be easier to treat gut microbes than to make genetic changes. And he is also excited about their potential ability to predict disease. Perhaps someday, changes in someone’s gut bugs could be indicators of impending illness – allowing shifts in diet or medications to restore the microbial balance before it leads to a serious health problem.
But while the data pour in, so begin the debates. One of the main discussions at the recent meeting concerned a study published last year, which suggested that people fall into three categories, or enterotypes as they are known, depending on the dominant group of gut bacteria living there. The idea of having a bug version of blood type that could predict a person’s risk of disease is a compelling one, but follow-up results presented in Paris suggest that the boundaries between types might be fuzzier than first imagined.
Ehrlich says the only way around that problem is to study many thousands of people. “Numbers count,” he says, adding that researchers in Europe, China, the United States and elsewhere need to collaborate to make sure their sample sizes are large enough to reflect meaningful differences in the microbiome. This also means researchers around the world need to develop protocols so they are all studying things the same way and their results are comparable.
Tracking the human microbiome also involves manipulating more data than scientists have ever dealt with before. Until recently, scientists had only been able to culture bacteria that could live in a petri dish; once they figured out how to separate the microbial from the human DNA, they began discovering dozens of new species. But they are collecting billions of bases, or gigabases, of DNA sequence data from complex populations – and figuring out which bits of DNA go together is only one part of the puzzle. How do you infer what those organisms are actually doing? How do they work together as an ecological system? How does this relate to human health – either normal or disease states – and how do you know whether the microbes have caused the condition, or whether they are just responding to the changing environment?
Curtis Huttenhower, assistant professor of Computational Biology and Bioinformatics at the Harvard School of Public Health, says the balance of microbes is more likely to matter than the individual strains. A few bad actors like salmonella will make you sick even in small concentrations, but for the most part, the good bacteria keep the bad in check, says Huttenhower, who studies inflammatory bowel disease. And many of us carry the bacteria Clostridium difficile around inside us all the time, with no ill effect. It is only after a shift in the microbe population – say, after a heavy-duty course of antibiotics that disrupts the balance – that the destructive power of C. difficile can be seen, causing symptoms ranging from diarrhoea to life-threatening inflammation of the colon.
Ruth, a college professor in suburban New York, knows this all too well. The 55-year-old, who asked that she be identified only by her first name, took a course of antibiotics in late 2006 for a bladder infection. A few months later, she needed antibiotics again, this time for a dental procedure – and that is when her problems began. She had terrible diarrhoea, and began losing weight and strength. Most troubling for a woman whose students still gave her credit for being “hot” on an online rating system: her hair started falling out in chunks. By May 2007 she was diagnosed with a C. difficile infection and started taking antibiotics to treat it. More than a full year later, she was still taking antibiotics – in much higher doses – and still getting sick every time she stopped.