hen we think about bad science, we tend to think about fields that rely on computer
models instead of evidence to make predictions. But there are hundreds, maybe thousands,
of falsehoods in science at any given time. Some of them are in fields like nutrition
and biological medicine, where a bad result can cost lives.
There are non-scientific myths, like the idea that vaccines cause autism, and scientific myths, which are falsehoods that are (for whatever reason) widely accepted by scientists. A scientific myth starts with one or two bad papers that tell us what we think ought to be true. Suddenly it becomes dogma, and anyone challenging the dogma is driven out of the field—not necessarily by being fired, but with papers going unpublished and grants going unfunded.
Cholesterol and ApoE
Many of our myths involve cholesterol. One myth is that it causes Alzheimer's disease. Apolipoprotein E (which everyone calls ApoE for short) is a protein that transports cholesterol to neurons. The myth says that ApoE is also involved in clearance of beta-amyloid. People buy some ApoE4, throw it on their cells, and conclude that the toxic effects account for the risk for ApoE4 in neurodegenerative diseases.
Here's the problem: ApoE never occurs in the unlipidated state. It's been known for over a decade that normal, lipidated ApoE doesn't bind beta-amyloid. Some researchers continue to use “naked” or unlipidated recombinant ApoE, which does bind beta-amyloid. Using recombinant ApoE is much easier, so the dogma persists, driving us in the wrong direction.
Bees using arithmetic
In recent years, a number of papers have shown up in top journals claiming that bees can count. This claim is based on training them to fly in response to a certain cue and drawing conclusions from their behavior.
Consciousness in bees, I'd believe. But counting is a formalized, procedural behavior that depends on language. Without language, there can be no counting, only vague perceptions of ‘more’ or ‘less.’ How do bees discriminate between more and less? We may never know, because the myth that they can count has gotten into the literature.
Blood-brain barrier antibodies
One time a famous researcher published a paper reporting how an antibody against transferrin, a protein that transports iron into the brain, could be used to carry DNA across the blood-brain barrier into the brain.
This was an important finding, but it turns out that many people regard it as an artifact. Apparently, they tried to replicate it with other antibodies and failed. But there's a simple explanation: if your antibody is too strong, it gets stuck to your target, so your drug, DNA molecule, or whatever, is never released into the brain. If your antibody is too weak, it doesn't bind to the target at all. Many academics never figured that out, and gave up on antibodies.
That didn't stop some big biotech companies from studying it. The pharma industry loves antibodies, mainly because they're more specific than old-fashioned small molecules, so the FDA is more likely to approve them. This is why so many of our new drugs are antibodies.
Ironically, then, much of the progress in this field is being made by big pharma. Unfortunately, industry scientists rarely publish their results, so most people don't know about it, and the science flounders.
Big pharma corporations, at least the ones that do research, aren't as evil as many people think. I'm not counting the ones that buy out their competitors, dump 10,000 experienced industry scientists onto the job market, and then run to the government complaining that they need more H-1Bs. There are good ones too, and for them the idea that they're in bed with the FDA is received with bitter laughter. Yet academic researchers are also held captive by the FDA, even if they don't realize it.
Protein kinase C
A good example of that is in protein kinase C, or PKC. A few years ago, I invented a bunch of new drugs called PKC epsilon activators. We planned to use them to treat Alzheimer's disease. It turned out that, though there was little effect on Alzheimer's, they have a strong effect on a different disorder for which existing drugs are inadequate.
So I wrote up a short grant and sent it to a foundation that was soliciting for preclinical projects. The reviewers loved the grant. They praised the writing. They commented on how well-designed the experiments were and how well it matched their funding goals. Then they rejected it.
It turns out there's a myth out there that PKC epsilon is an oncogene. It's simply not true. I've read many of the papers that claim it, and I'm utterly baffled by their reasoning. I even wrote an article debunking it. I (and others) did protein structural studies demonstrating it. I cited all this in the grant, to no avail.
If I were a 70-year-old with dementia, I would not care one whit about a one in a hundred or even one in ten chance that twenty years later, when I'm ninety, I might get cancer. But the press is powerful, and the FDA is almost totally risk-averse. With that falsehood in the literature, developing those drugs would be a waste of time.
Why does bad science get accepted?
We see a similar dynamics with glyphosate and baby powder, where a single bad paper is taken up by the press as a war cry against the hated chemical industry (typical scare headlines: 93% of Americans may have glyphosate in their bodies! Breakfast foods contain glyphosate, the weed-killing chemical linked to cancer!) The press wants you to think everyone in the industry is a Shkreli clone who suppresses harmful facts, raises prices, and poisons babies to boost profits.
I've written about this nonsense many times. High drug prices are where the free market could solve the problem easily. The high prices are mainly caused by the FDA's safety regulations. People rarely complain that there is too much safety regulation. People are extremely risk-averse, so to make drugs safe we have piles of regulations that raise the price not by two or three fold, but by a thousand fold. If this continues, some other big country whose name starts with a C will eventually take over, and another American industry will be lost. That is already well underway.
Other social pathologies are even harder to root out, because institutions have a vested interest in maintaining them. One particularly nasty pathology comes from giving absolute power to university bureaucrats. At some schools they're free to kill careers and destroy research projects at will, and so they do.
Those vicious bureaucrats get their power because of the way the National Institutes of Health and other funding agencies handle grants. It's enormously destructive to science. Before I moved to academia, I had no idea how bad it was.
Ethical judgments in science
Biologists often evaluate the importance of their work by how many lives will be saved as a result of their discoveries. I suppose they get this habit from the fact that the NIH has as its purpose, and even its very name, the improvement of human health.
This is essentially an ethical argument. It's tempting to conclude that standing in the way of curing some fatal disease is unethical. But there's a drawback of reducing ethics to biology. If someday all disease were eliminated, our sense of ethical values would collapse. This might be why so many people today invoke Hitler in trivial contexts: in an era of relative peace, they crave the moral clarity that was once found by opposing a ruthless and destructive enemy.
Some might say we should celebrate the bureaucrats who hold back science and sabotage careers. As religious people say, just as shadows prove the existence of light, bureaucrats prove the existence of wise human beings. And no doubt bureaucrats think we are the shadows.
When science is impaired by falsehoods, we can blame ourselves for not being skeptical enough or convincing enough. It's not a deliberate act, but a failing. But when some bureaucrat at a university decides on a whim to do something that prevents us from saving thousands of lives, it's infuriating, and it's hard not to think that it's also unethical.
dec 31 2018, 5:40 am. last edited jan 02 2019, 5:27 am
