The fatal flaw of the microbiome studies highlights the trap in science of the pursuit of the wild type human

The pursuit of “normalcy” has seriously led some scientists astray and there is no better example than the recent microbiome studies that have drawn great attention in the media. I found one aspect of their study design profoundly disturbing.

First to my core question.

Is there even such as thing as a normal person?

As a kid, if I asked such a question my somewhat evil older brothers would have pointed to me and yelled “NO!”

But seriously, are some people more normal than others?

wild t-shirt

Is there such a thing as a “control” person to whom others can be compared?

As we might say in science, is there  a “wild type” person?

(and no I don’t mean a partier, but rather one who is a normal reference to compare other abnormal people to).

For example, such a control person would have cells (their own and that of their microbiome) that are more normal than a random person like you and me.

Does they exist?

I don’t think so.

Why am I asking this?

Because in science we like to compare groups of things to a so-called “control” group that is often referred to as “wild type”…as opposed to mutant or otherwise abnormal. This is a basic, yet powerful approach for science, but it can also be a trap.

As I said above, a perfect example comes from the fascinating microbiome studies published recently to much fanfare. In fact, it’s been such a big week or so for the microbiome as Gina Kolata of the NYT described in an excellent piece that some of the devil seems to have got lost in the details. (For reference: (1) A beautiful chart can be found here at NYT showing the human microbiome. (2) Carl Zimmer has written a nice piece in the same place on what this all means…or at least what some of this massive effort means.)

One particular quote from Kolata’s excellent piece stood out (emphasis mine):

The first problem was finding completely healthy people for the study. The investigators recruited 600 subjects, ages 18 to 40, poking and prodding them. They brought in dentists to probe their gums, looking for gum disease, and pick at their teeth, looking for cavities. They brought in gynecologists to examine the women to see if they had yeast infections. They examined skin and tonsils and nasal cavities. They made sure the subjects were not too fat and not too thin. Even though those who volunteered thought they filled the bill, half were rejected because they were not completely healthy. And 80 percent of those who were eventually accepted first had to have gum disease or cavities treated by a dentist.

Oh, this is very bad!

The scientists not only pursued a false goal of “completely healthy people” but even worse they manipulated their subjects via dental work before studying them.

What the heck?

The scientists driving the project reportedly screened potential subjects and excluded those who had any sign of disease or other abnormality. Some of the subgroup of 600 who made the first cut and were chosen for potential microbiome analysis were then forced to have dental work to, I suppose, shift them more towards the control, wild-type “normal” human state.

In the end of the 242 subjects chosen for the microbiome are in my opinion not any more wild type than say the 242 people chosen at random from the NY city subways.

In fact, I bet that in reality that they are less “wild type” because they are too normal. They are freakishly normal.

Extending the logic of the scientists “cleaning” their subjects’ teeth before studying them, one might say that the only true “wild type” for such a study would be a human being who spent their life in a sterile bubble with no micro-organisms at all…but then of course such a person could not exist and would have no microbiome. I’m using this extreme hypothetical example to illustrate the point that in my opinion the scientists driving these studies have artificially drawn a line in the sand that basically says some people are more wild type when it comes to the bacteria that live on and in them, while others arbitrarily are essentially abnormal. I don’t believe science knows enough to make such a distinction. Further, the scientists manipulated some people in order to make them fit into the group on the other side of the line.

I’m not trying to bug the bug genomicists specifically, but rather more generally I’m making the point that we do not really know what people should be used as controls or considered “wild type” in science.

So, why pretend?

Certainly we should not change our subjects to try to make them “wild type”, right?

The microbiome design is just a great example of scientists falling into this trap.

The fact that 80% of the subjects eventually enrolled in the microbiome study had dental work first before being allowed to participate greatly concerns me….gee, don’t you imagine that changed the results? Fewer bacteria in the mouth, more in the blood stream perhaps and lodging in other organs?

Seriously, if 80% of the subjects had to have significant bacteria-removing dental work before they were subsequently studied perhaps that should tell you that the scientists should NOT have had the patients’ teeth cleaned or cavities filled at all because that 80% figure suggests that most people have issues with their teeth.

In fact most people have at least a mild form of gum disease if you are a stickler for dental hygeine.

Even ignoring the dental work itself, the fact that scientists pre-studied the hell out of their subjects via doctors testing them and poking and prodding these people like crazy could have introduced bias into the study as well.

And, no, I’m not missing the point.

I realize the scientists involved were aiming to get a snapshot of the microbiome that normally occupies the human body and to a large extent I am guessing they succeeded, but in the pursuit of normality I believe they likely skewed their own results dramatically based on subject bias (I’m not sure this is the right technical term, but you know what I mean) and even worse they manipulated their subjects via the dental work to make them acquire what the scientists perceived as a more wild type state.

That kind of thing is a big no-no in science, right?

This would be akin to me treating my lab’s stem cells with an epigenetic drug to try to make them more “normal” and then calling such cells the “controls”. Uh, no way!

Don’t get me wrong, I am a big fan of  projects like the microbiome and I don’t question their worth as some others including excellent writer Matthew Herper has done regarding the microbiome here. And sure we all need control subjects for our studies whether they are people, cells, or platypuses, but how we choose those controls in large part determines the outcomes of our studies.

I believe that as we go about our science that we have to be extracareful that in the name of pursuing “control” or “wild type” or “normal” subjects that we don’t fool ourselves into thinking that if we are studying people, that such people really exist.

…and if such people do exist, I don’t believe we know enough to properly identify them now anyway.

And certainly we should not manipulate them in an attempt to normalize them according to some preconceived notion that we have about normalcy.