Heritable Disease Hypotheses: Occam vs. Murphy

Riki's picture

Heritable Disease Hypotheses: Occam vs. Murphy

Two decades ago, the National Institutes of Health began the Human Genome Project, whose main goal was to sequence the entire human genome. The project was completed in 2003, but only several new disease treatments have arisen. Many scientists believed knowing the sequence would allow them to correlate genes with diseases. The Common Disease-Common Variant hypothesis has been used consistently to guide genomics research. According to this hypothesis, common major illnesses arise from inherited common alleles [1]. Therefore, the idea was that if scientists could figure out which common variants were linked to each common disease, they could develop pharmaceutical treatments for all common diseases. This was a reasonable assumption 20 years ago but no longer makes sense in light of current understandings. An alternate hypothesis is gaining attention, that of rare variants. Is one hypothesis better than the other, and if so, why?

 How can diseases like cancer be so widespread if they do not promote the survival of the species? The Common Disease-Common Variant hypothesis is helpful for explaining how diseases are so prevalent despite selective forces acting for the fittest population. The common diseases addressed by this hypothesis are those that act later in life, such as dementia. The variants responsible for these diseases can be inherited because the diseases do not affect reproduction since they present themselves after reproductive age [1]. Due to the pervasiveness of common diseases around the globe, it can be assumed that the associated common variants must have been in the genome for tens of thousands of years, when Homo sapiens were still confined to Africa [6].

Based on this hypothesis, scientists hoped to locate disease-related variants in the genome by using single-nucleotide polymorphisms (SNPs) as markers [1]. First, they had to throw millions of dollars into the HapMap Project, the goal of which was to determine common SNPs. The results of the HapMap and the Human Genome Project have been used in many genome-wide association studies (GWAS) to statistically correlate SNPs with diseases [1]. So why haven’t treatments been developed for all diseases?

It is becoming more and more apparent that the Common Disease-Common Variant hypothesis is not helpful. Though common variants can certainly contribute to disease, there are many observations the hypothesis does not account for.

One obvious problem is heritable early-onset diseases. If the disease presents before reproductive age, selective forces act against associated mutations/variants, meaning that the variants could not date back tens of thousands of years, let alone several generations, and that the variants are therefore not common but rare [2].

Another problem is that results of GWAS fail to predict genetic risk for disease. Furthermore, SNPs are only important if they are located within a portion of the genome that regulates genes or codes for proteins; however, less than 5% of the genome is involved in these functions [6], meaning the vast majority of SNPs have no biological relevance to understanding diseases [2]. This again leads to a new explanation, one of rare variants being largely responsible for diseases.

Indeed, the rare variant hypothesis is proving to be a more useful model for describing heritable complex diseases. It states that common variants rarely have severe effects, and that most complex diseases are the collective result of many rare variants [2]. Of course, research under this hypothesis is much more difficult and expensive – individual genomes have to be sequenced and compared to healthy controls.

In the past year, it has been suggested that scientists must consider the magnitude of genetic heterogeneity in disease. The failure of the HapMap and the dearth of new treatments since the completion of the human genome is a wake up call to scientists worldwide – human disease isn’t a four-piece puzzle! Human disease is incredibly complicated, and scientists should have been wiser than to delude not only themselves, but the international public as well, into believing that detection, treatment, and prevention of virtually all human disease was in the near future. The same disease can be caused by both rare and common variants [2]. And the same variant can lead to different diseases or no disease at all in different individuals based not only on what other variants are involved, but epigenetics, environmental factors, RNA regulatory elements, and dose and timing of gene expression [2].

An example of the complexity and heterogeneity of disease is inherited predisposition to breast cancer. So far, studies have determined that germline mutations in 13 genes are related to the susceptibility [2]. More genes are believed to be involved due to the amount of related individuals with breast cancer that lack mutations in the other 13 genes that have been studied [2]. In one of the most frequently mutated genes, BRCA1, there are over 1,000 different alleles that contribute to breast and ovarian cancer susceptibility [2].

If looking at disease from the rare variant viewpoint, the notion of disease treatment and prevention is bleak. This is probably one of the reasons scientists have only recently been giving the hypothesis serious attention – the task is too daunting. Does this explanation mean that treatment must be tailored to each individual? Not necessarily. Just because an individual has many rare variants does not mean the variants are all functional or causally linked to a disease. And just because many rare variants can be involved in one disease does not mean they are of equal effect. It is likely that mutations have varying effects, some more severe than others. What is still important and common to disease research is to determine which pathways and processes are critically affected by the disease, then to design drugs to target these pathways and processes [2].

Scientists hold onto the Common Disease-Common Variant hypothesis because it is a simple explanation that is easier to work with, which is basically an Occam's razor argument [5]. Many scientists agree that the HapMap was a waste of money, but that it may have been necessary to rule out the common variant hypothesis before proceeding down a different path [1]. It seems likely that is easier to get funding for research based on this hypothesis because it provides answers fairly quickly.

The rare variant hypothesis is not simple. It introduces Murphy’s law to the study of heritable diseases. Murphy’s law says that if something can go wrong, it will. In terms of genetic variation, any point mutation that doesn’t lead to spontaneous abortion is probably present in someone’s genome [2]. Since the world population and the number of base pairs in the human genome both rank in the billions, it seems possible that you could have a variant that no one else on the planet has. This sort of research requires massive amounts of funding, because it is very expensive to sequence individuals. Immediate answers are not guaranteed.

It is not helpful for scientists to waste resources fighting over which hypothesis is “right” or “wrong”. Both arguments can be made and empirically proven. And if both contribute to the current understanding of how the genome relates to disease, then neither method-inspired research should be disregarded. Even if scientists don’t quite know what to do with the human genome just yet due to technological deficits and general confusion, they should still be applauded for the new observations that have resulted from the Human Genome Project. It would have been disappointing to those who always ask questions if the solution to all diseases were found in the genome. Instead, more questions were unearthed such as: Why is over 95% of the genome junk DNA? If humans are so complex, why is the human genome no larger than those of “lower” organisms? Where is the line between normal and disease, particularly in the brain?

The Human Genome Project should be hailed as an increasingly rare way to do science – science based on discovery rather than hypothesis. Hypotheses are limiting, which is helpful in some contexts but not others. Unfortunately, when scientists begin researching a series of hypotheses, they can often get tunnel vision and fail to think about the issue in alternative ways. Research to produce new observations is more difficult to fund than research to prove existing observations, despite the ability of each to promote human understanding and knowledge. In today’s society, it seems that scientific research is a means to gain status and money rather than to further knowledge. Therefore, I suspect culture plays a large role in scientific funding. I don’t doubt that many studies are funded to assuage the general public. Which sounds more promising to the near future, an Occam’s razor approach to research or a Murphy’s law approach? In the context of disease research, Occam’s razor sounds like it would yield more immediate results, which is probably why the HapMap was funded, along with GWAS. Although the idea of matching up a few genes per disease sounds stupid now, scientists could not have known otherwise without first trying it. Scientific exploration is about trying and failing and trying again, all the while making observations and forming new understandings.

 

 

References:

[1] Revolution Postponed, Scientific American, October, 2010

[2] Genetic Heterogeneity in Human Disease, Cell, 16 April 2010

[3] A Decade Later, Genetic Map Yields Few New Cures, New York Times, 12 June, 2010

[4] Disease Cause Is Pinpointed With Genome, New York Times, 10 March, 2010

      [5] Genes, evolution, science education, and science, Serendip, 28 September, 2010

[6] How We Are Evolving, Scientific American, October, 2010

 

 

 

Comments

Paul Grobstein's picture

the genome project: implications for studying genes and disease

Interesting critique and prospectus.  Well worth extending to more specific examples, both of disease and of forms of research/therapies.  And thinking about public policy implications, past and future.
 
 

Post new comment

The content of this field is kept private and will not be shown publicly.
randomness