Update: A problem with the study that I describe in this post has surfaced within the scientific community, and has even received mass-media attention. It’s a technical problem of exactly the sort I warn about in the post. The only thing that’s unusual here is that the problem has been called out prominently.
A new study has shown that living into your nineties and beyond is a matter of having the right genes, according to recent headlines. The technique that the researchers employed to identify the genes of long life is much the same one behind many of the discoveries of links between genes and traits or genes and disorders that we hear about frequently. To oversimplify a bit, the game involves looking at many genes in a large group of people who have some trait in common, and then trying to spot which particular genes turn up more frequently in these folks than in a large group people who don’t have the trait.
I recommend not paying much attention to this study, or to most studies that purport to have identified key genes behind some important trait or disorder, and for a few reasons. First of all, this basic gene-linking technique, although an impressive one and getting better, is still beset with a variety of inherent limitations and potential flaws that often lead to simply misidentifying genes as being closely linked to a trait in most people when they are not, and misses many and possibly most of the genes that actually influence the trait. Part of the problem is that the technique is a highly statistical one based on probabilities involving all sorts of manipulations that can leave the picture convoluted and murky. In addition, the very notion that a relatively small number of genes is likely to play a huge role in determining a trait or disease risk is for the most part a mistaken one, or at best a gross oversimplification. Genes act not alone or even as a small band, but rather as a vast orchestra. Even more confusing, genes can be “turned off” so that they might as well not be there, or partly turned on so that they contribute only weakly to a trait or disease risk. What turns genes on and off? Each other, the environment, other molecules in our cells, and strips of DNA that aren’t even parts of genes. It gets worse: Genes can appear in single form, or in multiple copies that increase their influence in ways most gene tests don’t detect. They can pop out of their slot in the genome, replicate themselves, and then reinsert themselves somewhere else in the genome. They can flip around backwards, which also isn’t detected by gene tests, even though a backwards gene can behave differently. Throw in the fact that genes can be changed by viruses, and that any protein made by a particular gene can do different things in different people, or different things in the same person at different times depending on what else is going on in the body, and you’ve got complexity of such staggering breadth and depth that scientists who are being honest about the situation admit they’ll be lucky to make a small dent in the task of sorting it all out in their lifetimes.
Consider: A 2009 study of about 6,000 people came up with a technique for predicting the height of a person based on looking at the 54 genes most closely linked to height. The results turned out to be one-tenth as accurate as averaging the heights of both parents and adjusting for sex, a technique introduced in 1886 by statistician Sir Francis Galton. University College Dublin biomolecular researcher Helen Colhoun and colleagues have estimated that 95 percent of studies that find gene links are just plain wrong. Harvard genetics researcher Joel Hirschhorn and colleagues surveyed the medical literature on 166 published genetic links to illness that were each examined in at least two other published studies, and found that only six of the links–less than four percent–held up in all the studies. John Ioannidis of Tufts University Medical School has calculated the average odds of a gene-link study being right at one out of hundreds, or worse, depending on the disease. In the case of schizophrenia, for example, Ioannidis reckons a one out of 2,000 chance of a gene-link study being right. His estimates suggest you would have almost as good a chance of identifying some types of genuine gene links by throwing darts at a diagram of the relevant sections of the human genome as you would by reading research journals. Researchers claim their hit rate has improved quite a bit in the past few years, but we won’t know the extent to which that’s true for a while to come. Needless to say, all these substantial uncertainties and vagaries often get swept under the carpet in the reported findings of an exciting new study, and are almost always ignored in the mass-media reports.
None of this is to say that the study in question here didn’t correctly identify many genes that may in fact in some cases be potentially important contributors to living a very long life. But even if the claim is largely or at least partly right, it’s still a misleading one. When we hear it, we assume that we’ll need many or most of these genes to have a good shot at a long, healthy life–and that’s an absolutely unwarranted conclusion. People who live extraordinarily long lives need everything to go in their favor, including many things that may be largely irrelevant to those of us who hope to keep decent health “merely” into our 80s. Assuming we need the genes common in people who live to be 100 may be like assuming a good-performing car needs to have the components common to race cars.
Even if it does indeed help quite a bit to have the right genes in order to live a long and healthy life, it also helps to get a lot of other things right. What has come through from all the genetic studies more clearly than anything else is that for most of us genes really don’t influence your destiny as much as, or at least any more than, your behavior does. Folks, if you want to live relatively long lives, and perhaps more importantly stay relatively healthy late into that life, there is powerful evidence that eating sensibly, exercising, breathing clean air, avoiding undue stress, and refraining from obviously bad habits such as smoking, heavy drinking and not wearing seatbelts, can for many and possibly most of us trump genes. In fact, for all we know, some of the genes that the very-long-lived had in common in this study were genes that influence healthy behaviors–in which case anyone engaging in the behaviors would get the same benefits, whether they had the genes or not.
Your fate is probably not genetically sealed–it is much more likely that your health is to a large extent under your control. Don’t let announcements of marvelous gene-link breakthroughs lead you to abdicate that responsibility.
David H. Freedman 