The development of the brain involves many thousands of different gene products acting in hundreds of distinct molecular and cellular processes, all tightly coordinated in space and time – from patterning and proliferation to cell migration, axon guidance, synapse formation and many others. Large numbers of proteins are involved in the biochemical pathways and networks underlying each cell biological process. Each of these systems has evolved not just to do a particular job, but to do it robustly – to make sure this process happens even in the face of diverse challenges.

Robustness is an emergent and highly adaptive property of complex systems that can be selected for in response to particular pressures. These include extrinsic factors, such as variability in temperature, supply of nutrients, etc., but also intrinsic factors. A major source of intrinsic variation is noise in gene expression – random fluctuations in the levels of all proteins in all cells. These fluctuations arise due to the probabilistic nature of gene transcription – whether a messenger RNA is actively being made from a gene at any particular moment. The system must be able to deal with these fluctuations and it can be argued that the noise in the system actually acts as a buffer. If the system only worked within a narrow operating range for each component then it would be very vulnerable to failure of any single part.

Natural selection will therefore favour system architectures that are more robust to environmental and intrinsic variation. In the process, such systems also indirectly become robust to the other major source of variation – mutations.

Many individual components can be deleted entirely with no discernible effect on the system (which is why looking exhaustively for a phenotype in mouse mutants can be so frustrating – many gene knockouts are irritatingly normal). You could say that if the knockout of a gene does not affect a particular process, that that means the gene product is not actually involved in that process, but that is not always the case. One can often show that a protein is involved biochemically and even that the system is sensitive to changes in the level of that protein – increased expression can often cause a phenotype even when loss-of-function manipulations do not.

Direct evidence for robustness of neurodevelopmental systems comes from examples of genetic background effects on phenotypes caused by specific mutations. While many components of the system can be deleted without effect, others do cause a clear phenotype when mutated. However, such phenotypes are often modified by the genetic background. This is commonly seen in mouse experiments, for example, where the effect of a mutation may vary widely when it is crossed into various inbred strains. The implication is that there are some genetic differences between strains that by themselves have no effect on the phenotype, but that are clearly involved in the system or process, as they strongly modify the effect of another mutation.

How is this relevant to understanding so-called complex disorders? There are two schools of thought on the genetic architecture of these conditions. One considers the symptoms of, say, autism or schizophrenia or epilepsy as the consequence of mutation in any one of a very large number of distinct genes. This is the scenario for intellectual disability, for example, and also for many other conditions like inherited blindness or deafness. There are hundreds of distinct mutations that can result in these symptoms. The mutations in these cases are almost always ones that have a dramatic effect on the level or function of the encoded protein.

The other model is that complex disorders arise, in many cases, due to the combined effects of a very large number of common polymorphisms – these are bases in the genome where the sequence is variable in the population (e.g., there might be an “A” in some people but a “G” in others). The human genome contains millions of such sites and many consider the specific combination of variants that each person inherits at these sites to be the most important determinant of their phenotype. (I disagree, especially when it comes to disease). The idea for disorders such as schizophrenia is that at many of these sites (perhaps thousands of them), one of the variants may predispose slightly to the illness. Each one has an almost negligible effect alone, but if you are unlucky enough to inherit a lot of them, then the system might be pushed over the level of burden that it can tolerate, into a pathogenic state.

These are the two most extreme positions – there are also many models that incorporate effects of both rare mutations and common polymorphisms. Models incorporating common variants as modifiers of the effects of rare mutations make a lot of biological sense. What I want to consider here is the model that the disease is caused in some individuals purely by the combined effects of hundreds or thousands of common variants (without what I call a “proper mutation”).

Ironically, robustness has been invoked by both proponents and opponents of this idea. I have argued that neurodevelopmental systems should be robust to the combined effects of many variants that have only very tiny effects on protein expression or function (which is the case for most common variants). This is precisely because the system has evolved to buffer fluctuations in many components all the time. In addition to being an intrinsic, passive property of the architecture of developmental networks, robustness is also actively promoted through homeostatic feedback loops, which can maintain optimal performance in the face of variations, by regulating the levels of other components to compensate. The effects of such variants should therefore NOT be cumulative – they should be absorbed by the system. (In fact, you could argue that a certain level of noise in the system is a “design feature” because it enables this buffering).

Others have argued precisely the opposite – that robustness permits cryptic genetic variation to accumulate in populations. Cryptic genetic variation has no effect in the context in which it arises (allowing it to escape selection) but, in another context – say in a different environment, or a different genetic background – can have a large effect. This is exactly what robustness allows to happen – indeed, the fact that cryptic genetic variation exists provides some of the best evidence that we have that the systems are robust as it shows directly that mutations in some components are tolerated in most contexts. But is there any evidence that such cryptic variation comprises hundreds or thousands of common variants?

To be fair, proving that is the case would be very difficult. You could argue from animal breeding experiments that the continuing response to selection of many traits means that there must be a vast pool of genetic variation that can affect them, which can be cumulatively enriched by selective breeding, almost ad infinitum. However, new mutations are known to make at least some contribution to this continued response to selection. In addition, in most cases where the genetics of such continuously distributed traits have been unpicked (by identifying the specific factors contributing to strain differences for example) they come down to perhaps tens of loci showing very strong and complex epistatic interactions (1, 2, 3). Thus, just because variation in a trait is multigenic, does not mean it is affected by mutations of small individual effect – an effectively continuous distribution can emerge due to very complex epistatic interactions between a fairly small number of mutations which have surprisingly large effects in isolation.

(I would be keen to hear of any examples showing real polygenicity on the level of hundreds or thousands of variants).

In the case of genetic modifiers of specific mutations – say, where a mutation causes a very different phenotype in different mouse strains – most of the effects that have been identified have been mapped to one or a small number of mutations which have no effect by themselves, but which strongly modify the phenotype caused by another mutation.

These and other findings suggest that (i) cryptic genetic variation relevant to disease is certainly likely to exist and to have important effects on phenotype, but that (ii) such genetic background effects can most likely be ascribed to one, several, or perhaps tens of mutations, as opposed to hundreds or thousands of common polymorphisms.

This is already too long, but it begs the question: if neurodevelopmental systems are so robust, then why do we ever get neurodevelopmental disease? The paradox of systems that are generally robust is that they may be quite vulnerable to large variation in a specific subset of components. Why specific types of genes are in this set, while others can be completely deleted without effect, is the big question. More on that in a subsequent post…

The problem with this question is that it limits the search space for the solution. It forces our thinking further and further along a certain path, when what we really need is to draw back and question the assumptions on which the whole approach is founded. Rather than asking what is the right answer to this question, we should be asking: what is the right question?

The idea of performing genome-wide association studies for complex disorders rests on a number of very fundamental and very big assumptions. These are explored in a recent article I wrote for Genome Biology (referenced below; reprints available on request). They are:

1) That what we call complex disorders are unitary conditions. That is, clinical categories like schizophrenia or diabetes or asthma are each a single disease and it is appropriate to investigate them by lumping together everyone in the population who has such a diagnosis – allowing us to calculate things like heritability and relative risks. Such population-based figures are only informative if all patients with these symptoms really have a common etiology.

2) That the underlying genetic architecture is polygenic – i.e., the disease arises in each individual due to toxic combinations of many genetic variants that are individually segregating at high frequency in the population (i.e., “common variants”).

3) That, despite the observed dramatic discontinuities in actual risk for the disease across the population, there is some underlying quantitative trait called “liability” that is normally distributed in the population. If a person’s load of risk variants exceeds some threshold of liability, then disease arises.

All of these assumptions typically go unquestioned – often unmentioned, in fact – yet there is no evidence that any of them is valid. In fact, the more you step back and look at them with an objective eye, the more outlandish they seem, even from first principles.

First, what reason is there to think that there is only one route to the symptoms observed in any particular complex disorder? We know there are lots of ways, genetically speaking, to cause mental retardation or blindness or deafness – why should this not also be the case for psychosis or seizures or poor blood sugar regulation? If the clinical diagnosis of a specific disorder is based on superficial criteria, as is especially the case for psychiatric disorders, then this assumption is unlikely to hold.

Second, the idea that common variants could contribute significantly to disease runs up against the effects of natural selection pretty quickly – variants that cause disease get selected against and are therefore rare. You can propose models of balancing selection (where a specific variant is beneficial in some genomic contexts and harmful in others), but there is no evidence that this mechanism is widespread. In general, the more arcane your model has to become to accommodate contradictory evidence, the more inclined you should be to question the initial premise.

Third, the idea that common disorders (where people either are or are not affected) really can be treated as quantitative traits (with a smooth distribution in the population, as with height) is really, truly bizarre. The history of this idea can be traced back to early geneticists, but it was popularised by Douglas Falconer, the godfather of quantitative genetics (he literally wrote the book).

In an attempt to demonstrate the relevance of quantitative genetics to the study of human disease, Falconer came up with a nifty solution. Even though disease states are typically all-or-nothing, and even though the actual risk of disease is clearly very discontinuously distributed in the population (dramatically higher in relatives of affecteds, for example), he claimed that it was reasonable to assume that there was something called the underlying liability to the disorder that was actually continuously distributed. This could be converted to a discontinuous distribution by further assuming that only individuals whose burden of genetic variants passed an imagined threshold actually got the disease. To transform discontinuous incidence data (mean rates of disease in various groups, such as people with different levels of genetic relatedness to affected individuals) into mean liability on a continuous scale, it was necessary to further assume that this liability was normally distributed in the population. The corollary is that liability is affected by many genetic variants, each of small effect. Q.E.D.

This model – simply declared by fiat – forms the mathematical basis for most GWAS analyses and for simulations regarding proportions of heritability explained by combinations of genetic variants (e.g., the recent paper from Eric Lander’s group). To me, it is an extraordinary claim, which you would think would require extraordinary evidence to be accepted. Despite the fact that it has no evidence to support it and fundamentally makes no biological sense (see Genome Biology article for more on that), it goes largely unquestioned and unchallenged.

In the cold light of day, the most fundamental assumptions underlying population-based approaches to investigate the genetics of “complex disorders” can be seen to be flawed, unsupported and, in my opinion, clearly invalid. More importantly, there is now lots of direct evidence that complex disorders like schizophrenia or autism or epilepsy are really umbrella terms, reflecting common symptoms associated with large numbers of distinct genetic conditions. More and more mutations causing such conditions are being identified all the time, thanks to genomic array and next generation sequencing approaches.

Different individuals and families will have very rare, sometimes even unique mutations. In some cases, it will be possible to identify specific single mutations as clearly causal; in others, it may require a combination of two or three. There is clear evidence for a very wide range of genetic etiologies leading to the same symptoms. It is time for the field to assimilate this paradigm shift and stop analysing the data in population-based terms. Rather than asking how much of the genetic variance across the population can be currently explained (a question that is nonsensical if the disorder is not a unitary condition), we should be asking about causes of disease in individuals:

- How many cases can currently be explained (by the mutations so far identified)?

- Why are the mutations not completely penetrant?

- What factors contribute to the variable phenotypic expression in different individuals carrying the same mutation?

- What are the biological functions of the genes involved and what are the consequences of their disruption?

- Why do so many different mutations give rise to the same phenotypes?

- Why are specific symptoms like psychosis or seizures or social withdrawal such common outcomes?

These are the questions that will get us to the underlying biology.

Mitchell, K. (2012). What is complex about complex disorders? Genome Biology, 13 (1) DOI: 10.1186/gb-2012-13-1-237

Manolio, T., Collins, F., Cox, N., Goldstein, D., Hindorff, L., Hunter, D., McCarthy, M., Ramos, E., Cardon, L., Chakravarti, A., Cho, J., Guttmacher, A., Kong, A., Kruglyak, L., Mardis, E., Rotimi, C., Slatkin, M., Valle, D., Whittemore, A., Boehnke, M., Clark, A., Eichler, E., Gibson, G., Haines, J., Mackay, T., McCarroll, S., & Visscher, P. (2009). Finding the missing heritability of complex diseases Nature, 461 (7265), 747-753 DOI: 10.1038/nature08494

Zuk, O., Hechter, E., Sunyaev, S., & Lander, E. (2012). The mystery of missing heritability: Genetic interactions create phantom heritability Proceedings of the National Academy of Sciences, 109 (4), 1193-1198 DOI: 10.1073/pnas.1119675109