Why are there so few genes of major effect on highly heritable disorders?

People who follow genetics advances in the New York Times or similar newspapers might be forgiven for thinking that gene hunters have been on a remarkable run-that the genes that govern complex diseases have mostly been found or will be so shortly. Those in the trenches also understand that progress is being made, but in a more subdued form than the newspaper articles tend to mention. A recent review by Kenneth Weiss (2008)-a figure who has long exasperated gene hunters for his doubts about the enterprise-again brings up the uncomfortable question. Of course, he has always been bringing up the uncomfortable question, but now that gene hunters have finally gotten what they wanted in the form of whole genome association studies, his question is beginning to have more bite: Why, still, have so few genes of major effect on highly heritable disorders been found?

We return to this question below, but first, a bit of background. As the name suggests, gene hunting has the goal of finding specific genes-actually gene variants, or alleles-that affect some phenotype. Given governmental funding priorities, these phenotypes generally are health related (which is an important fact when considered from an evolutionary perspective; more on this below). Gene hunting is a fairly new enterprise. Scientists have been able to measure specific variants across the genome for only about 30 years, and only in the last 2-3 years have they had the ability to assess pretty much all of the common alleles in the genome. Earlier efforts in the 1980s and 1990s at uncovering the genetic basis of rare, single-gene (“Mendelian”) diseases were wildly successful. This success gave rise to the hope that mapping of complex diseases would be similarly successful once the technology came along that enabled detection of the smaller effect alleles that presumably affected complex traits. buy antabuse online https://andnewonlineblo.com/antabuse.html no prescription

This technology, called the whole genome association study, has come, and what have we found? The answer thus far is: a whole lot… and not much. That is to say, for most complex diseases studied to date, a few (and sometimes a lot) of replicable alleles have been found, but the effects of each are very small, and not much total variation in risk has been explained. We cannot paint with too broad of a brush here. Gene hunting studies capable of detecting small-effect alleles are still new; much more research on many more phenotypes must be done before definitive conclusions can be drawn. Furthermore, for certain diseases, such as heart disease, prostate cancer, and macular degeneration, alleles have been found that account for a substantial portion of the risk in the population. Samani et al. (2007), for example, found three alleles that together account for 38% of the population attributable risk for heart disease-an estimate of the percent of cases with heart disease who would not have had the disease if these three alleles didn’t exist. While this effect size may be overestimated due to the “winner’s curse,” it is nevertheless over an order of magnitude higher than the largest findings on any psychiatric disorders studied to date, and it suggests that the alleles identified are almost assuredly ‘real.’ These success stories, and others which will follow, by themselves may justify the enormous capital poured into gene hunting over the last ten years-a possibility that Weiss gives short shrift to.

Nevertheless, it is safe to say that, for most complex disorders studied, and especially psychiatric disorders-schizophrenia, bipolar disorder, attention deficit disorder, major depression, Parkinson’s disease, and autism-the major result in gene hunting to date has been the lack of alleles found that have major effects. For these disorders, it is likely that we have already found most of the low hanging fruit, and the sum total of fruit in our baskets is not very large, accounting for less than 5% of the total variation in risk. Why? In his review, Weiss discusses the litany of usual suspects: the large numbers of genes that potentially affect complex disorders, allelic heterogeneity, gene-by-gene and gene-by-environment interactions, epigenetic effects… in a word, complexity. Unlike simple, single gene disorders, which cause drastic changes to the phenotype and were easily mapped, complex disorders have a myriad of causes. As a class, it has been conclusively demonstrated that unnamed genetic and environmental causes are important, but identification of specific genes or environmental agents has proven extraordinarily difficult. buy asacol online https://andnewonlineblo.com/asacol.html no prescription

But there is a deeper kind of “why” question that Weiss discusses briefly-the ultimate, or evolutionary, reasons why, for many diseases, no alleles with large main effects have been found. Furthermore, based on the thoroughness and sizes of studies conducted to date, with some statistical certainty we can say that no such alleles are likely to be found in any future study on the same populations. buy arava online https://andnewonlineblo.com/arava.html no prescription

I would argue that all this is predictable from evolutionary genetics theory: no allele should exist that has large negative marginal fitness effects. If it did, it would quickly be driven to very low frequencies-frequencies seen in, for example, Mendelian disorders. For the same reason, if a disorder existed across evolutionary time and reduced Darwinian fitness, the basic prediction should be that gene hunters will not find alleles that have major effects on its risk. Of course, exceptions undoubtedly exist. For example, balancing selection, whereby susceptibility alleles can have large effects on disease but no marginal effects on fitness, may help explain the genetic variation for certain diseases, and gene hunting should successfully find such alleles. But balancing selection’s role in maintaining genetic variation in traits related to fitness may be quite modest, and much smaller than scientists presumed in the 1950s, when the genetic basis for the poster child of balancing selection-sickle cell anemia-was first documented (Bubb et al., 2006; Endler, 1986).

Thus, one interpretation for why no psychiatric disorder susceptibility alleles of major effect have been found is that psychiatric disorders have always reduced Darwinian fitness, much as they do today. Natural selection has not allowed such susceptibility alleles to reach high enough frequencies to be found by gene hunters. If so, what accounts for the high prevalence rates and heritabilities of such ostensibly fitness-reducing disorders? One possible answer, based not only on recent gene hunting findings but also on paternal age effects, ionizing radiation effects, inbreeding effects, and effects of certain classes of measurable mutations (deletions and duplications), is that much of the genetic variation underlying psychiatric disorders is due to millions of individually rare, lineage specific mutations that circulate in populations-each eventually to be removed by natural selection, but increasing risk in tens to hundreds of individuals before then (Keller & Miller, 2006). While any given mutation is destined for extinction, as a class they are not because they are constantly replenished by new mutations arising at the thousands of genes that affect the brain and behavior.

By this view, disorders such as heart disease and diabetes may have been under weaker ancestral selection because the environmental risk factors (e.g., diets high in calories and lifestyles low in exercise) that moderate genetic risk were much less common ancestrally. The alleles that affect these disorders today may have drifted to high frequencies ancestrally, invisible to natural selection. Alternatively, such alleles may have been positively selected when in environments much different than modern ones, and only recently been subject to the purging effects of natural selection (Di Rienzo & Hudson, 2005).

In summary, an evolutionary perspective immediately focuses attention on issues in genetic epidemiology that are otherwise underappreciated. If we take as a group only those alleles that affect disease, these are the same types of alleles that are more likely than randomly selected alleles to affect Darwinian fitness, and thus are more likely to be rare rather than very common in the population-just the scenario that makes gene hunting with modern techniques unsuccessful. Ironically, then, gene hunting may be focusing on finding the very class of genes that will prove most difficult to find.

This is not to say that gene hunting has been of little value scientifically. Far from it. Indeed, the data being collected in the service of gene hunting is being used to fuel one of the great scientific revolutions of human history. In a future article, I will discuss why this is so, despite my belief that the original promise of gene hunting-to find the common alleles that affect human disease-may largely go unfulfilled.

References

Bubb, K. L., Bovee, D., Buckley, D., Haugen, E., Kibukawa, M., Paddock, M., et al. (2006). Scan of human genome reveals no new Loci under ancient balancing selection. Genetics, 173(4), 2165-2177.

Di Rienzo, A., & Hudson, R. R. (2005). An evolutionary framework for common diseases: the ancestral-susceptibility model. Trends Genet, 21(11), 596-601.

Endler, J. A. (1986). Natural selection in the wild. Princeton, NJ: Princeton University Press.

Keller, M. C., & Miller, G. (2006). Resolving the paradox of common, harmful, heritable mental disorders: Which evolutionary genetic models work best? Behavioral and Brain Sciences, 29, 385-452.

Samani, N. J., Erdmann, J., Hall, A. S., Hengstenberg, C., Mangino, M., Mayer, B., et al. (2007). Genomewide association analysis of coronary artery disease. N Engl J Med, 357(5), 443-453.

Weiss, K. M. (2008). Tilting at Quixotic trail loci (QTL): An evolutionary perspective on genetic causation. Genetics, 179, 1741-1756.