When it comes to science, I have the patience of a rabid fox, trapped in a cage, in front of which a wounded rabbit is standing. My family, the folks in my lab and the need for sleep balance this nascent madness. But sometimes the caged fox of mania escapes; sometimes when everyone else sleeps I can’t resist the run.
Image 0. A young fox pretending it isn't thinking about a rabbit. Image courtesy Rob Lee.
Today was one of those days. We saw another glimpse into the life inside belly buttons. Belly buttons are ridiculous and yet the life we study in them is not; it includes both dangerous and life saving species though in just what mix and why, well, that is what we’d like to know. As a result we have, over the last few years, worked with more than 500 people to sample the life in their belly buttons. It has not always been pretty (Imagine emails from concerned and infected citizens that include photos. Yes, we see those. No, please don’t send them.), but the aim was to have a consistent part of the body through which we might understand the differences more generally, person to person, in what lives on skin.
Your skin is covered in life, a fine featheriness of single-celled organisms, your verdant cloak of existence, a cloak so woven into your existence that it is not clear where it ends and you begin. This is life that colonizes you during or before birth and accumulates through living. It also sheds. The more you sit in a place, the more that place is filled with the microbes from your skin. We are all like Pigpen, tracking a cloud of our microbes wherever we go.
But which life? And can we understand what determines which life? That is our rabbit-hopping quarry. Or it was. In route we stumbled into something else and ran after it down another trail. That something else was the realization that, at least in our first datasets, if you knew which species were abundant (present in many cells on the average person) and frequent (present on many people) in one group of folks you could predict the same for another group of folks, and perhaps any other you examined. In other words, some bacteria are predictably abundant and frequent AND those bacteria appear to come from just a small subset of the many lineages of life, an unusually small subset, a handful of sun-shaped life forms in a universe filled with stars. To us this was a big deal because thousands of species live on skin, but if only a minority is predictably important, knowing the story of the life of the skin might actually be tractable.
This is what we found, but here is where I tell you about a thing scientists know but don’t share with the public. I think the formal term for it is soul-crushing self-doubt. Scientists are taught to be skeptical. They are taught that most of what they learn about the world, most of what we know about the world, is wrong. They are taught to poke relentlessly at our existing knowledge, to look for weakness, errors, things into which the finger sinks deeply. What wakes them up at night is the sense that the thing they have just discovered will be found out to be the thing that is wrong.
The only recourse for this crushing, scientific self-doubt is, of course, to repeat observations and experiments, or to cry. So in the mornings after sleepless nights, we do more science to check, again and again, our ideas hoping that if something is wrong or not quite right that we will find it out before our peers do. The only thing worse than realizing you have totally misunderstood the gears behind the clock of life is having one of your colleagues realize it for you. Colleagues are, of course, not the same as friends.
If an idea stands the test of this additional prodding then, it just might be true. (The doubt never disappears, it just diminishes, just as science never reaches the truth, it just approaches it ever more closely). It is for this reason that what happened to me tonight, on one of those reckless evenings of plunging headlong into data was so exciting.
We recently got our first glimpse at the data from the last 284 people who sampled their own belly buttons. The process works more or less like this. People sample their belly buttons. We extract DNA from their samples (which takes more time than you think), we amplify that DNA (which takes more time than you think), we find the money to pay for that DNA to be sequenced (which takes more time than you think), we then send that DNA off to a company that does the sequencing, the DNA then comes back in an email as data (after more time than you might think). Someone then processes those data to turn them from zillions of nucleotides into something that can be jammed into a spreadsheet or statistical software or other programs (which takes WAY more time than you might think) and then and only then we have a data matrix you might recognize, one with the kinds of bacteria in the columns and the identity of the belly buttons in the rows. With that matrix, we can begin to see our quarry.
It so happens, that thanks to the work of many people, including the citizens who sampled their belly buttons, the students who processed those samples, the postdocs who coordinated the processing, Holly Menninger who made sure the bill was paid, and another postdoc who processed the data further, we now have a data matrix. And so if you sent us a sample of your belly button and are waiting to hear what lives in your belly button, we will tell you very, very, soon (It also takes WAY more time than you might think to write 284 emails plus the emails to folks whose samples didn’t work). But before we do, I am going to tell you what I peeked at. I went through and with these new data I looked to see whether what we found for the first data, those that we already published on, held. It was not what we proposed to study next, it was not the big question, but it was the wounded rabbit in front of my fox cage, the thing on which I could seize to see for myself whether my sleepless nights were deserved. I wanted to know if, as in our first samples, the common and frequent species were predictably common and frequent and the uncommon and infrequent species were predictably uncommon and infrequent. It might, for a million reasons, have been otherwise. Part of me was sure it would be.
I should have gone to bed. It was late when I first saw the data. I had things to do the next day. I had a morning to wake up to, but I could not resist and so I split the data into, for a starting point, three random hunks of the same number of people and examined the ability of the frequency of different kinds of bacteria in each hunk to predict that in each other hunk.
Here are the results as I first saw them. You should think like a scientist and, as a starting point, assume that I was wrong. But I was not.
Image 1. The number of people on which bacterial taxa occur as a function as their total number of occurrences. Species that are common overall are common in each random subset of the data.
In short, it seems as though across very different people, our rule of commons and uncommons, the law of oligarchs as we have called it elsewhere, holds. The most frequent species are taxa of Corynebacterium and Staphylococcus (which are also among the most frequent bodily taxa we find in houses). This is lovely for a million reasons, but it is most lovely tonight because I poked my finger hard into what we discovered and my finger didn’t sink in. The idea held against the pressure, which doesn’t mean it was right, but it certainly means, it is not yet wrong.
Image 2. One (crude) measure of the relative abundance of different kinds of bacteria in our most recent 274 samples. Number on the x-axis is a person. The y-axis shows the relative number of reads of each kind of bacteria. On many people, a Corynebacterium is, by this metric, the most abundant kind of bacteria in the belly button, but a Staphylococcus is also very common. The third most common kind of bacteria does not even have a genus name.
Now we get to explore the question that started us off in the first place, but we do not need to study all of the species. The predictable presence of the common, frequent species means that they are the big story, what we most need to explain. It is late tonight, but I may go try now to figure them out. How could I not? The rabbit waits a little further down the trail.