The New Eugenics – same as the Old Eugenics?

Did I miss a memo? Has eugenics somehow become respectable again?

Source

There has been a lot of talk lately, in the blogosphere at least (1, 2, 3, 4, 5), about the idea of using molecular genetics to predict and select for higher intelligence in humans (through pre-implantation screening of embryos, for example). The prevailing view among many discussing this idea seems to be that if we can do it, we obviously should do it. The casualness with which this conclusion is reached is astonishing to me, given the history of humanity’s efforts in this area. To many commentators, it seems to be a given that having more intelligent people, across the population, is not only obviously a good thing, but one that supersedes any other considerations.

Selecting for increased intelligence doesn’t sound so bad, when you phrase it like that, until you realise that it actually involves the converse – selecting against individuals with lower predicted intelligence. I am not ascribing the following chain of thought to any particular persons, but here is the fundamental logic of eugenics, applied to intelligence:

For any individual, being more intelligent is better than being less intelligent. (All else being equal, that’s fair enough, I suppose). People who are more intelligent are therefore better than people who are less intelligent. (See how easy it is to get there?) At least, it would be good if we had more of the former and less of the latter. We should, as a society, seek ways to ensure that is the case. In the past, this would have involved policies on who is allowed to live or breed or migrate into a society, or inducements to get the more clever people to breed like they vote in Chicago – early and often. Nowadays, if we can employ pre-implantation genetic screening to predict intelligence, then we should use that method, or at least make it available, to select and implant those embryos that are predicted to be more intelligent. This will inevitably be at the expense of ones predicted to be less intelligent. The former should be granted life and the latter should not.

Is all that just self-evident? Is that how we should define progress in our society?

The amazing thing, in the pieces I have been reading recently, is that something approaching this position seems to have been reached not after lengthy and sober consideration of the moral and ethical issues surrounding the idea, but in total disregard for them. The following questions don’t seem to have come up: Is it right to claim some people are superior to others or of “higher quality”? Is it right to actively select between embryos (or to selectively abort foetuses) on any criterion? (Many people would say no, though it already happens routinely for serious medical conditions, and even for sex in many parts of the world). If there are some criteria that can be considered legitimate, what are they? How do we decide? Who makes those decisions? Should society as a whole ever have the right to dictate such decisions? Or should society allow complete freedom to individuals to make such decisions on any criteria they wish? If selection is permissible, is intelligence really the primary trait on which such selection should be based? What about kindness or decency or bravery or empathy or not being a douche? Do any of those get a look in? Would we lose anything from human society by selecting purely for those who perform better on IQ tests?

The impression one gets is that the people proposing such ideas think the world would be a better place if there were more people like them in it. The spectacle of cosseted academics bemoaning the degraded intellect of the masses and suggested something should be done about it is not an appealing one. And it is not without consequences.

There seems to be little recognition of the potential harm to the reputation of genetics as a science when it is associated with public claims of this sort. This discipline still bears the taint of previous misuses, most notably as justification for the murderous eugenic policies of Nazi Germany or enforced sterilisations of the “feeble-minded” in many US states which ran from the early 1900’s to as late as 1977 in North Carolina. Many other countries enacted similar policies.

Source

The themes of genetic classism and discrimination and of elitist scientists “playing God” resonate widely in our culture (from Shelley’s Frankenstein to GATTACA to the X-Men). Indeed, the extensive coverage of a study on the genetics of IQ that is currently underway at the Beijing Genomics Institute (BGI) suggests that the media knows a good story when it sees it. It seems to me that this has attracted attention not because of any scientific advance or discovery (the study has not yet been completed) but because of the way those involved and commenting on it have acted as cheerleaders for the idea of prenatal prediction and selection.

Here’s Prof. Geoffrey Miller (of NYU), in an interview for an article egregiously entitled “China is engineering genius babies” on vice.com (whatever that is):

“How does Western research in genetics compare to China’s?
We’re pretty far behind. We have the same technical capabilities, the same statistical capabilities to analyze the data, but they’re collecting the data on a much larger scale and seem to be capable of transforming the scientific findings into government policy and consumer genetic testing much more easily than we are. Technically and scientifically we could be doing this, but we’re not.”

Some would argue it is not the place of scientists to decide the ethical issues – it is our job just to do the science. If society abuses it, well, that is not our fault. This is a case where I strongly disagree – we cannot disentangle the moral issues from the scientific ones. It is too easy to use scientific findings to justify policies that would otherwise be deemed abhorrent; too easy, as Hume noted, to mistakenly derive a prescription of how things ought to be from a description of how they are.

In this case the science is too complex and our understanding still far too fragmentary to even describe how things are. But reading some of the commentaries one would think that our ability to predict intelligence based on molecular genetics is really just around the corner; that we will have this knowledge in hand within a few years and Pandora’s box will have been opened, whether we like it or not. I find this scenario highly implausible, for several reasons.

First, we have not yet identified any genes “for intelligence”. We know many that, when mutated, can cause intellectual disability (many hundreds, in fact), but none that clearly contribute to variation in the normal range (normal in the statistical sense of that word). Zero, zip, bupkis. We are starting from effectively complete ignorance as of this moment. In fact, we don’t even understand the genetic architecture of intelligence. It is clearly very highly heritable, but we don’t know how many genes are involved, either across the population or in any individual, we don’t know whether the genetic variants are common or rare, we don’t know whether they specifically affect intelligence or have more general effects on robustness of the genetic program and its execution to build an efficient brain and we don’t know how multiple such variants would interact with each other. That’s a lot of don’t knows.

The answers to those questions will determine the best strategies for finding variants that affect intelligence and also, crucially, our ability to predict an individual’s IQ based on signatures that we can only detect by averaging across the population. If we want to be fanciful, we can imagine a future scenario where we have in fact identified many genetic variants across the population that clearly contribute to differences in intelligence. Some may be common, but my expectation is that most would be quite rare. Now we want to look at some new individual’s DNA and predict their IQ based on that knowledge (or maybe look at two individuals and predict which one’s IQ will be higher, even if we can’t put a number on it).

Here are the problems: first, IQ is indeed highly heritable, but a lot of the variation across the population is non-genetic (at least 20-30%); that imposes a significant limit on accuracy of even a perfect genetic predictor. Second, if IQ is largely affected by rare mutations, then each new person will have some IQ-affecting variants that we have never seen before in our population sample and that we will be unable to recognise as such. Third, any individual will also have a unique, never-before-seen combination of variants, which may interact in highly unexpected ways. Finally, any such predictor would have to be extremely precise to distinguish between the IQ of not just any two random individuals, but two siblings, where the range will obviously be much narrower.

To paraphrase Yogi Berra, making predictions is hard, especially about the future. But I am willing to go out on a very sturdy limb and predict that we will not be able to build useful predictors for IQ any time soon. We’re not there, we’re not nearly there and there may even be fundamental limits that mean we will never get there.

 

A post from 2012 by Greg Cochran goes even further, suggesting that a variety of approaches to improve intelligence are imminent, from selection to molecular interventions designed to correct mutations lowering intelligence. This not only fails to consider any of the ethical and moral issues described above, it similarly ignores the additional ones that arise when considering modifying the human germline! It also greatly exaggerates our technical abilities to do that. Yes, we can modify the germline in model organisms like mice, but what this simple statement glosses over is the fact that generating any such genetically modified individual involves a lot of trial and error. This science is messy. Most of the embryos (or cells) one tries to modify do not get modified in the expected way and one has to screen through many hundreds typically to get ones with the desired change. (Even those can sometimes have other, random changes one didn’t plan for). This is clearly not a strategy we could countenance in humans.

In the meantime, before we go proposing scientifically impractical and morally questionable extreme measures, we have a proven and powerful tool to make people smarter: education.

The genetics of emergent phenotypes

Source

Why are some brain disorders so common? Schizophrenia, autism and epilepsy each affect about 1% of the world’s population, over their lifetimes. Why are the specific phenotypes associated with those conditions so frequent? More generally, why do particular phenotypes exist at all? What constrains or determines the types of phenotypes we observe, out of all the variations we could conceive of? Why does a system like the brain fail in particular ways when the genetic program is messed with? Here, I consider how the difference between “concrete” and “emergent” properties of the brain may provide an explanation, or at least a useful conceptual framework

There is now compelling evidence that disorders like epilepsy, schizophrenia and autism can be caused by mutations in any of a very large number of different genes (sometimes singly, sometimes in combinations). This is fundamentally changing the way we think about these disorders. It is no longer tenable to consider them as unitary categories. Instead, it is very clear that the underlying etiology is extremely heterogeneous – possibly more so than for any other human disease.

How can this fact be explained? Why is it that mutations in so many different genes (perhaps thousands) can give rise to the specific phenotypes associated with those disorders?

The normal logic of genetic analysis entails some correspondence between the phenotypes associated with mutations in specific genes and the functions of the products encoded by those genes. This connection between mutation and phenotype is one of the main reasons why experimental genetics is so powerful. For example, if we carry out a genetic screen for mutations affecting cell death in a worm, or embryonic patterning in a fruit fly, the expectation is that the genes we discover will be directly involved in those processes. That is how the molecular processes regulating cell death and embryonic patterning were discovered.

This logic can sometimes be applied to humans too – but not always. Let’s consider two genetic conditions – microcephaly and epilepsy – both affecting the brain, but in quite distinct ways.

MRI of child with microcephaly (top). Source.

Microcephaly is a rare condition characterised by a small brain. In particular, the cerebral cortex is smaller than normal, due to a defect in the generation of the normal number of neurons in this brain area. It can be inherited in a simple, Mendelian fashion, due to a mutation in any one of at least six different genes. Remarkably, the proteins encoded by these genes are all involved in some aspect of cell division of neuronal progenitors. In particular, they determine whether early divisions expand the initial pool of progenitors (in the normal situation) or prematurely generate neurons (when any of these genes is mutated). 

The genes implicated in microcephaly are thus directly involved in the process affected: the generation of neurons in the cerebral cortex. It is not too inaccurate to say that that is what these genes are “for”.

This is not the case for epilepsy. It too can be inherited due to specific mutations, but there are many, many more of them and the known genes involved have diverse functions: from controlling cell migration or specifying synaptic connectivity to encoding ion channels or metabolic enzymes. These are not genes “for” regulating the spatial and temporal dynamics of electrical activity in neuronal networks.

Source

Put another way, the reason that we see microcephaly as a phenotype is that there are genes that control the process we are looking at – generation of neurons in the cortex. The existence of that phenotype thus reflects a property of the genetic system. In contrast, the generation of seizures does not relate in any meaningful way to the genetic system – instead, it is an emergent property of the neural system. We see that phenotype not because there are many genes directly controlling that process, but because it is a state that the brain tends to get into, in response to a wide diversity of insults. (Indeed, seizures are one of the symptoms sometimes associated with microcephaly).

I have used the term “emergent” twice now without defining it and had better do so before I get pilloried by those allergic to the word. There is good reason for a negative reaction, as the term is fraught with multiple meanings and seemingly mystical connotations.

Concepts of emergence range from the mundane (the whole is more than the sum of its parts) to the magical (where the behaviour of a system is not reducible to or predictable from the state and interactions of all its components, and where new properties emerge apparently “for free”). In fact, it is possible to allow for new principles and properties at higher levels without invoking such mystical concepts or over-riding the fundamental laws of physics.

Nature is organised hierarchically into systems at different levels. Subatomic particles are arranged as atoms, atoms into molecules, molecules in cells, cells into tissues and organs, and ultimately organisms, individual organisms in collectives and societies. At each level, qualitatively novel properties arise from the collective action of the components at the level below. Emergence refers to the idea that many of these properties are highly unexpected and extremely difficult to predict (though not necessarily impossible in principle). One objection to the term is that it is therefore essentially a statement about us (about our level of understanding) and not about the system itself. I think it goes further than that, however, and does denote some principles of nature that actually exist in the world, regardless of whether we understand them or not.

While the emergent behaviour of a system is reducible to the microstates of the components at the level below and the fundamental physical laws controlling them, the emergent properties are not deducible purely from those laws. To put it another way, the microstates of a system are sufficient to explain the properties or macrostates observed at any moment but are not sufficient to answer another question – why those properties exist. Why is it that those are the properties observed in that particular system, or that tend to be observed across diverse systems? These properties arise because additional laws or principles apply at the higher level, which constrain the arrangements of the components at the lower level to some purpose.

Many of these principles of functional organisation are abstract and apply to diverse systems – principles of network organisation, cybernetics and control theory, information content, storage and processing, and many others. All of these principles constrain the architecture of a system in a way that ensures its optimality for some function.

In artificial design of complex machines, these engineering principles are incorporated to ensure that the parts are arranged so as to produce the desired functions of the system as a whole. In living organisms, it is natural selection that does this work, leading to the illusion of design (or teleonomy), apparent only in hindsight. System architectures that produce useful emergent properties at the higher level (i.e., the phenotype of the organism, which is all that selection can see) are retained and those that do not are removed. In this way, the abstract engineering principles constrain the functional organisation of the components of the system – there are only certain types of arrangements that can generate specific functions. This is top-down causation, but over a vastly different timescale from the mystical, moment-to-moment versions proposed by some emergence theorists.

Let’s move from the abstract to a more specific example and think about how these issues relate to the kinds of phenotypes we see when a system is challenged. Consider a complicated, highly specified system like a fighter jet. It has many different parts – engines, turbines, fuselage, flaps, wheels, weapons, etc. – each with multiple subcomponents and each with a specific job to do. If we were examining multiple designs for a jet, we might consider various specs for, say, the turbines. We might vary the number of blades, their size, angle, etc. These are all concrete properties of the system and there are a finite number of them.

Source: Newcastle University

Contrast that with an emergent property of the jet, something like aerodynamic stability, fuel efficiency or even something harder to define, like “performance”. These properties depend on the specs of all the individual components of the plane, but also, more importantly, on their functional organisation and the interactions between them (and the interactions of the whole system with the environment). A property like performance is not easily linked to any specific component – instead it emerges in a highly non-linear fashion from the specs of all of the components of the system and how they are combined.

If you randomly broke one component in the jet, it is thus much more likely that you would affect performance than that you would affect the turbines specifically. The bits of the turbines are not “for performance”, per se – they are for whatever job they do in the turbine. There aren’t any bits of the jet that you would say are “for performance”, in fact, but all of them can affect performance.

The kinds of functions affected by disorders like epilepsy, autism or schizophrenia are like performance. For epilepsy, it is the highest-order properties of neural systems – the temporal and spatial dynamics of electrical activity. For schizophrenia and autism, it is functions like perception, cognition, sense of self, executive planning, social cognition and orderly thought – the most sophisticated and integrative functions of the human mind. These rely on the intact functioning of neural microcircuits in many different areas and the coordinated actions of distributed brain systems. Evolution has crafted a complex and powerful machine with remarkable capabilities, but those capabilities are consequently vulnerable to attack on any of a very large number of components.

Thinking about these phenotypes in this way thus provides an explanation for why epilepsy and schizophrenia are so much more common than microcephaly. The mutational target – the number of genes in which mutations can cause a particular phenotype – is much, much bigger. (This obviates the need to invoke some kind of counter-balancing benefit of the mutations that cause these disorders to explain why they persist at a high frequency. The individual causal mutations do not persist – they are strongly selected against, but new mutations arise all the time. Under this mutation-selection balance model, the prevalence of a disorder is determined by an equilibrium between the mutational target size and the strength of selection).

But this perspective does not explain everything that needs explaining. These conditions do not manifest simply as a general decrease in brain “performance”. It is not just that normal brain functions are somewhat degraded. Instead, qualitatively new states or phenotypes emerge. Psychosis is probably the most striking example – psychiatrists call the hallucinations and delusions that characterise psychosis “positive symptoms”, reflecting the fact that they are a novel, additional manifestation, not just a decrease in the function of specific mental faculties (as with the negative symptoms, such as a decrease in working memory).

Why does this specific, qualitatively novel state arise as a consequence of so many distinct mutations? This is where our fighter jet runs out of steam, as a (now mixed) metaphor. The problem with that metaphor is that fighter jets are designed and built from a blueprint. Parts of the blueprint correspond to parts of the jet and their arrangement is also specified directly on the blueprint.

This is not at all the case for the anatomy of the brain. The genome is not a blueprint – there are no parts of the DNA sequence that correspond to parts of the brain. Instead, the structure of the brain emerges through epigenesis – the execution of the developmental algorithms encoded in the genome, which direct the unfolding of the organism. (Aristotle coined this term epigenesis, which contrasted with the prevailing theory, known as pre-formationism – the idea that the fertilised egg already contains within it a teeny-weeny person, with all its bits in place, which simply grows over the period of gestation).

The ultimate phenotype of an organism is thus emergent in the more common sense of that word – it is something that arises over time. This emphasises the need to consider developmental trajectories when trying to understand the highly heterogeneous etiology of these disorders.

Modified from: Kitano, 2004

Complex, dynamic systems tend to gravitate towards certain stable patterns of activity and interactions in the network. Such patterns are called “basins of attraction” or “attractors”, for short. You can think about them like hollows in a flat sheet, with the current network state represented by the position of a ball rolling over this landscape. The flat bits of this landscape represent unstable, fluid states that are likely to change. The hollows represent more stable states – particular patterns of activity of the network that are easy to get into and hard to get out of. Generally speaking, the deepest such basin will represent the typical pattern of brain physiology. It takes a big push to get the ball up and out of this basin. But there are other basins – alternative stable states and the pathophysiological state we recognise as psychosis may be one of those.

Such alternate states may exist as by-products of the functional organisation of the system. The system architecture will have been selected to robustly generate a particular functional outcome. However, when individual components are interfered with, new functional states may emerge – ones that are unexpected and that the system has not been selected to produce. They arise instead as an emergent property of the broken system, as a specific failure mode.

It is vital to understand not just the nature of such states, but the trajectories that dynamic systems (in this case organisms) follow to get into them. (In dynamic systems, the relations between components of the system are not fixed but change over time). If we take our flat sheet and tilt it from one end, turning it into a board with channels in it, rather than hollows, then we can represent the path of a developing organism through phenotype space, over time. 

This is Conrad Waddington’s famous “epigenetic landscape” – a powerful metaphor for understanding how dynamic systems can be channelled into specific, stable states. The shape of the landscape will be determined by an individual’s genotype – some people may have much deeper channels heading towards typical brain physiology while others may have a greater chance of heading towards particular pathophysiological states, like psychosis or epilepsy.

One reason why psychosis and epilepsy may be common states is that they can reinforce themselves, through altering the relations of components of the system. In a process known as “kindling”, seizures induce changes in neuronal networks that render them increasingly excitable and more likely to undergo further seizures. A similar dynamic process, involving homeostatic processes in dopaminergic signaling pathways, may be involved in psychosis. These homeostatic mechanisms in the developing brain can, under certain circumstances, be maladaptive, pushing the network state into a particular pathophysiological pattern, in response to diverse primary insults.

Finally, a developmental perspective can also provide an explanation for the high levels of phenotypic variability observed with mutations conferring risk for psychiatric disorders. Such mutations can manifest in different ways, statistically increasing risk for multiple conditions. A person’s risk for developing schizophrenia is statistically much higher if they have a close relative with the condition, but their risks of developing autism or epilepsy (or bipolar disorder or depression or attention-deficit hyperactivity disorder) are all also higher. Even monozygotic (“identical”) twins are often not concordant for these clinical diagnoses. So, while genetics can lead to a much greater susceptibility to these conditions, whether a specific individual actually develops them depends also on other factors.

One of those factors, often overlooked, is intrinsic developmental variation. The development of the brain is inherently probabilistic, not deterministic (more like a recipe than a blueprint). This is evident at the level of individual cells, nerve fibres and synapses and can manifest at the macro level as variation in specific traits or symptoms in individuals with the exact same genotype.

Waddington’s landscape can also visualise this important role of chance in determining an individual’s eventual phenotypic outcome. If you roll a marble down this board multiple times, you will get multiple outcomes, essentially by chance (due to thermodynamic noise at the molecular level, affecting gene expression, protein interactions, etc.).

For a concrete property such as brain size, the amount of noise affecting the phenotype will be low, as a small number of components and processes are involved. The correspondence between genotype and phenotype will therefore be quite linear for concrete properties. In contrast, emergent properties that depend on large numbers of components will be more subject to noise and the relationship between genotype and phenotype will be far less linear.  This explains why mutations causing psychiatric disorders show lower penetrance and higher variability in phenotypic expression – this is the predicted pattern for emergent properties.

To sum up, thinking about these kinds of disorders as affecting emergent properties can explain why they are common, why the genes responsible are so diverse, why their products are only distally and indirectly related to the processes affected by the clinical symptoms and why the phenotypic outcomes are inherently variable.

Genes, brains and human nature; the joys and challenges of writing about science for non-scientists.

This post first appeared as part of a SpotOn NYC special event on Communication and the Brain (March 2013), on nature.com.  

When I started the Wiring the Brain blog a few years ago, it was with the intention of writing mainly for students, scientists and clinicians in the fields of genetics and neuroscience. Many of the posts deal with advances in our scientific understanding of the causes of neurodevelopmental and psychiatric disorders. Perhaps because of that, or just due to general interest in the broader themes, the blog has also become widely read by non-specialists.

This presents some exciting opportunities to write in a different way and to convey the excitement of the field of neurogenetics to the general public, but also raises some particular challenges. The biggest difference I have found is that the assumption of a shared, global perspective is not always valid. When writing for scientific colleagues there is an implicit expectation of a common starting point – not just a background of specific knowledge about a subject, but a foundation of wider shared beliefs that do not need to be articulated explicitly.

In the context of the themes of the Wiring the Brain blog, these include: that the diversity of life arose through evolution by natural selection; that human genes and human brains are not that different from animal genes and animal brains; that human minds emerge from the activity of human brains and nothing else; that variation in genes can affect behaviour; that studying the components of a system is a good way to make progress in understanding the whole; and most fundamentally, that the scientific method is the best way we have of finding things out and not just one of many “ways of knowing”.

Being challenged on some of those positions has been an eye-opener. It makes you look for the evidence that supports them. For evolution, that is pretty much all the observations ever made in the field of biology. In that circumstance, the job becomes marshalling that evidence to convince someone who may not have heard it all laid out before.

On another topic, this prompted one of the few blogposts I have written that veered into philosophical territory, entitled “On discovering you’re an android”. This presented the overwhelming evidence for neuroscientific materialism, the position that the mind is what the brain does, with no need to invoke any immaterial or supernatural stuff. This theory is both counter-intuitive and highly discomfiting. After all, it doesn’t feel like you are an android (though one made out of meat). One’s “self” feels pretty real and stable and the idea that it emerges from and relies on the continued activity of the neurons in your brain can leave one feeling existentially precarious. It is human nature to recoil at this idea, but that’s what all the science says and science wins.

Or does it? Not everyone would accept that assertion. Many argue that science is just another belief system with no special claim to validity. Part of the point, for me, of writing the blog is to illustrate how science works, how we accumulate evidence, how current paradigms can be challenged, modified or even overturned by new data. That is, in fact, the polar opposite of a belief system. Also, it works!

A related claim, and a common enough reaction to writings on the subject of human nature is that scientists like myself are just reducing human existence to mindless biochemistry, even down to physics. This charge of “Reductionism!”, which comes from psychologists as much as from members of the general public, misses an important distinction between methodological and theoretical reductionism. Yes, geneticists approach a problem by looking for components of a system that can vary in a way that affects the performance of the system. That is an experimental approach that has proven hugely powerful, allowing one to identify important parts of a system and ultimately analyse how they function together to mediate that system’s functions. If the system is a human being, then that necessarily entails understanding it in the context of its relations to other human beings.

Changes in single genes can have large effects on behavioural traits in humans (for example, see here on genetic influences on impulsivity). This does not mean that the behaviour in question is mediated by a single gene. Nor does it mean that human behaviour is determined by genes – it simply says that variation in that component of the system can contribute to variation in patterns of behaviour over time. But no matter how precisely that sentence is worded, it is important to realise, as a writer, that it can still be misconstrued by people who are “reading between the lines”, or extrapolated to infer a much broader claim consistent with a reader’s preconceived notions of what scientists think.

That view may have been informed by the shorthand that many scientists and journalists use about “genes for this” and “genes for that”, which does indeed sound very deterministic and reductionist. The absurd hype in many press releases, driven by pressures for the next grant, adds greatly to this problem. (There’s no shortage of that kind of thing in coverage of neuroscience either, now affectionately known as “neurobollocks”). This kind of wording is sloppy, sensationalist and deeply wrong at a conceptual level. Still, “Scientists discover one of many factors that contributes to people’s behavioural tendencies, which express themselves over time in the context of each individual’s life experiences” does not make a good headline.

It is no wonder, then, that readers often conclude that a larger claim is being made – exposure to relentless hype in science coverage fully justifies that expectation. I occasionally get comments starting, “So, what you’re really saying is…”, which continue to say something I really wasn’t saying. Anticipating and pre-empting these kinds of over-extrapolations can be an important part of this kind of writing.

Another challenge, especially in writing about the causes of clinical disorders such as autism or schizophrenia, is that these issues are necessarily fraught for people suffering from these conditions or with children who are affected. Many have strongly held views about the causes of their or their child’s particular condition, sometimes unfortunately founded on misinformation. The scientific hoax linking autism with vaccines has been incredibly hard to dislodge from the public’s consciousness. It is almost impossible to combat moving personal anecdotes with dry statistical data showing no association. The former are much more psychologically available – we are cognitively wired to learn from specific instances of apparent correlations and very poorly adapted for statistical thinking. The apparent “autism epidemic” reinforces the notion of some environmental causes, though there is clear evidence that this reflects only an increase in awareness and diagnosis, not of the true underlying rates of the condition. Showing how such data can be evaluated, scientifically, can go some way to equipping people with the tools to distinguish solid claims from one-off observations and correlation from causation.

My experience of writing for scientists and non-scientists alike has been very enjoyable and stimulating and I have learned a lot from it. I think it has made me a better teacher and a more thoughtful researcher. I have been struck in particular by both the tremendous interest in science among the general public and by how poorly it is served by traditional media. Blogging provides an exciting opportunity for scientists to help fill that void directly.

Photo Credit: Oliver Burston, Wellcome Images - Position of the brain inside the head
Digital artwork/Computer graphic 2004 Collection: Wellcome Images
Copyrighted work available under Creative Commons by-nc-nd 2.0 UK: England & Wales, seehttp://images.wellcome.ac.uk/indexplus/page/Prices.html

The Trouble with Epigenetics (Part 2)

In Part 1 of this blog, I considered the various definitions of the term epigenetics and the confusion that can arise when they are conflated. Molecular epigenetic mechanisms modify chromatin structure and provide a means to stabilize a particular profile of gene expression. They also allow that profile to be passed on to a cell’s descendants, through mitosis. For this reason, epigenetic profiles have been called “heritable” (meaning through cell division). It is easy to see how that definition can be extrapolated to the idea that epigenetics could provide a means of heredity from one generation to the next.

This idea has attracted substantial interest, with many people seeming to think it overturns classical genetics (the inheritance of characters based on DNA sequence), and rehabilitates Lamarckian evolution by supplying a respectable molecular mechanism. This view has gained particular prominence of late in the study of behaviour and psychiatry, with the proposal that transgenerational epigenetic inheritance can provide a mechanism of heredity that explains the so-called “missing heritability” of psychiatric disorders.

The idea of transgenerational epigenetic inheritance is that epigenetic marks laid down in the cells of one generation (in response to some environmental factor or experience) can be stably passed through meiosis (into the germ cells) and thus affect some traits in the next generation. This kind of thing is indeed known to happen in some very specific circumstances, which are highly illustrative. This review by Daxinger and Whitelaw gives an excellent, up-to-date synthesis of this field. Most of the known examples involve the establishment of specific chromatin structures at DNA repeats or transposable elements – i.e., it occurs in very particular genomic contexts. In many cases, the transmission of this chromatin state through the gametes depends on an RNA molecule, as opposed to the more traditional DNA or histone protein modifications.

This is a fascinating area of biology (though more an embellishment than an overthrowing of normal mechanisms of inheritance), but is it relevant to psychiatric disorders? In particular, can it contribute to the heritability of such conditions?

Twin and family studies have clearly shown that many psychiatric disorders are highly heritable (with h2 values around 65-70% for schizophrenia and 75-85% for autism). Nevertheless, large-scale studies aimed at detecting DNA differences that contribute to this heritability have not turned up much. At least, this is true for genome-wide association studies (GWAS), which look for differences in frequency of common genetic variants between large numbers of cases and controls. Some people are interpreting the failure to identify specific causal variants as implying that the traits are really not that genetic after all. This is a complete fallacy.

GWAS analyse only the parts of the genome that harbor a common variant or single-nucleotide polymorphism (SNP) – these are positions in the DNA sequence where two forms commonly exist in the population (some might have an “A” base, while others might have a “T” in that position, for example). For autism, large-scale GWAS have not found any replicable SNPs associated with the disease. For schizophrenia, recent (still unpublished) very large GWAS have reportedly found 62 replicated SNP associations, but collectively these still only explain ~3% of the heritability. Does this mean that the observed heritability is really not accounted for by variation in DNA sequence? Not at all.

It has become clear over the last few years that rare mutations make a very large contribution to individual phenotypes, especially to the occurrence of diseases. GWAS do not survey these rare mutations and their failure to fully account for the heritability of the disorders therefore means nothing - really nothing at all - regarding that heritability. These disorders are still just as heritable and that heritability still means that most of the variance in whether people get the disease or not is down to genetic differences (in the DNA sequence). We do not need epigenetics to come to the rescue here. Unless rare mutations are also exhaustively surveyed and found to be unable to collectively account for the observed heritability, there is nothing to explain.

More to the point, even if there were, transgenerational epigenetic inheritance could not explain it. The heritability of these disorders has been estimated mainly from twin studies – these show that monozygotic twins are much more phenotypically similar than dizygotic twins. As the twins in each case share the same uterine and family environment, we can conclude that the reason MZ twins are more similar phenotypically to each other is because they are more similar genetically. The heritability of a trait or a disorder can be estimated from the strength of this effect and is defined as the proportion of phenotypic variance across the population that can be attributed to genetic variance. So, unless the supposed epigenetic marks affect MZ twins more consistently than DZ twins (and there’s no reason why they should), this mechanism provides no explanation for the key observation. Even if epigenetic mechanisms can provide some means of heredity from one generation to the next, that is not what heritability measures.

Moreover, the evidence that epigenetic mechanisms can provide a means of heredity for behavioural traits is not strong. In Part 1 of this blog I cited a few examples where particular experiences have a lasting effect on behaviour of an organism, in part by stably altering gene expression in particular cells in the brain through molecular epigenetic mechanisms. These kinds of effects can indeed be perpetuated across generations, for example, in the well-known observation that stressed female rats have stressed offspring. That is because stress reduces maternal care of the newborns, which is itself stressful and which sets up long-term changes in expression of the glucocorticoid receptor. But this is a behavioural transmission: mom’s behaviour affects offspring’s behaviour – repeat. This is not an example of epigenetic inheritance via the gametes, which is what has been proposed as a possibly important mechanism.

For that to happen, the epigenetic marks laid down in the brain by experience would have to also be laid down in the germ cells, maintained through the genomic “rebooting” that happens in the fertilized zygote (where the vast majority of epigenetic marks are wiped clean), carried through subsequent development, surviving the epigenetic upheavals entailed in the generation of all the embryonic cell-types that are ancestors of the eventual cells in the brain where the effect on this specific behaviour is mediated.

This is more an intuition than an argument, but this scenario seems inherently far-fetched to me. One expects experiences to modify gene expression in the brain, but not in the gametes. Scientists should, of course, be prepared to be surprised and delighted by unimagined discoveries that overturn our preconceptions. On the other hand, a healthy level of skepticism is usually a good idea, especially in cases where such discoveries are not attended by strong evidence.

So, is there any evidence that this can happen? Given the possible confounds attending maternal transmission, several groups have looked for evidence of transmission through the paternal germline. A study by Isabelle Mansuy and colleagues illustrates some of the problems that I see with this literature. This is definitively entitled “Epigenetic Transmission of the Impact of Early Stress Across Generations” and is cited over 100 times, so it has clearly been influential. This study involved stressing a young animal by unpredictably removing its mother for several hours at a time. When these animals grow up they show residual effects of this maltreatment (details below). So far, so good. It is further claimed, however, that this effect is passed on to the next generation and even to the subsequent one, through the male germline. Now, this is an extraordinary claim, one that should require extraordinary evidence. Instead, the bar seems to have been lowered.

I do not mean to pick on this one paper, but it exemplifies a general problem in this field – that of too many researcher degrees of freedom. This refers to studies that are exploratory in nature and that do not define a specific hypothesis to be tested in sufficient detail prior to collecting data. Researchers looking for a difference between two groups may carry out a range of tests and report any test that shows a difference or may decide, after the fact, to look for effects just in one sex or the other, or just in one age group, or just at one time-point, etc., etc. If there is no reason, a priori, to expect the effect to be specific in such a manner, then this is just significance-fishing. If the significance estimates are not corrected for the multiple tests carried out, then they do not accurately convey how surprised we should be by any one finding. (This is the difference between the odds of you winning the lottery and the odds of the lottery being won). See this xkcd cartoon for a great illustration.

The study by Mansuy and colleagues illustrates the cardinal sins of significance-fishing. The male mice that are directly stressed by having their mothers removed show “depressive-like” behaviours on two tests – the forced swim test and the sucrose preference test. These males were then bred to female animals that have not been stressed in any way and the behaviour of their offspring was tested. The result? Females, but not males in the next generation showed a significant difference (p < 0.1) on the forced swim test, but not the sucrose preference test. So, four tests were carried out and one was “significant”. In the next generation (breeding from what were phenotypically normal males), the pattern was reversed! – males showed a difference (p < 0.5, again, only on one test), while females showed none. (Additional tests of sensitivity to stress showed an effect in first-generation females but this time in second-generation females, while males showed no difference). None of these results was corrected for multiple testing, nor is there any putative mechanism or a priori hypothesis to explain the sex-specificity of the effects (which, to any impartial observer, seems like random noise).

Despite the weakness and selectivity of the actual data, the claim in the abstract of this paper is both forceful and sweeping: “Most of the behavioral alterations are further expressed by the offspring of males subjected to maternal separation”. This is clearly not supported by a proper statistical evaluation of the actual observations. Actually, I don’t know if I worded that strongly enough: the data in this paper do not support any conclusion of a behavioural effect being transmitted across generations. That’s better.

The same problems are evident in a recent paper claiming epigenetic transgenerational inheritance of a “cocaine-resistance” phenotype. In this case, it was expected that cocaine exposure in one generation would lead to increased sensitivity to it in subsequent generations. In fact, the reverse was found, and only in one sex. So, the direction of this effect was a surprise and presumably there would have still been a paper if the mice were more sensitive, rather than less. Similarly, there was no a priori expectation of a sex effect or hypothetical mechanism to explain it. If it had only shown up in females, I expect we would have heard about that too. That’s four bites of the statistical cherry.

Adding genomics to these studies (looking at profiles of gene expression or methylation, for example), and highlighting those genes that show a “significant” difference when considered alone, compounds this problem of multiple testing – the poor cherry is just being gnawed on now in the most unseemly fashion.

In general, the evidence of a real behavioural effect being transmitted through males to the next generation is not compelling. These studies also suffer from an additional possible confound – the possibility that interacting with a stressed or strung-out male animal will alter the behaviour of the female, post-mating, so that maternal care is also changed. This would be quite different from the model that some experience causes an epigenetic mark in the male germ cells that, in effect, transmits a “memory” of that experience to the next generation. The best way to test for such an effect is to see if it is really transmitted through the male gametes themselves using in vitro fertilization. One study that did just that found effectively no such transmission (again taking multiple tests into account).

So, while epigenetic mechanisms are implicated in the long-term effects of certain experiences, the evidence that such effects can be transmitted through the germline to subsequent generations is, to my mind at least, extremely weak. And even if they could be, they certainly cannot represent a solution to the mystery of the “missing heritability” for psychiatric disorders. These disorders are as heritable as they ever were and that still implicates differences in DNA sequence. Jut because we haven’t found them yet doesn’t mean we should start looking somewhere else. 

Because real genetics.

The Trouble with Epigenetics (Part 1)

“You keep using that word. I do not think it means what you think it means”. The insightful Inigo Montoya.

Epigenetics is a word that seems to have caught the public imagination. This is especially true among those, both in science and without, who decry what they see as genetic determinism or at least an overly “genocentric” point of view. Our genes are not our fate, because epigenetics! Such-and-such disorder is not really genetic, because epigenetics! Acquired characteristics can be inherited, because epigenetics!

The trouble with epigenetics is that the word means very different things in different contexts. Each of them may be quite valid, but when these meanings are conflated or when the intended meaning is not specified, the word becomes dangerously ambiguous. This is especially evident in the fields of behavioural and psychiatric research where the term is much abused, often, it seems to me, to give an air of mechanistic truthiness to ideas that are in reality both speculative and vague.

Originally coined by Conrad Waddington in his famous “epigenetic landscape”, the word signified the emergence of the eventual phenotype of an organism through the processes of development, starting from a particular genetic profile. It was derived from Aristotle’s term “epigenesis”, which means pretty much the same thing – that organisms emerge through a program of development, as opposed to the theory of preformationism (where a teeny organism is already formed inside an egg and simply grows). Waddington’s new term incorporated the idea of a genetic profile, which shapes the metaphorical landscape over which each individual developing organism travels, channeling them with greater or lesser probability toward certain outcomes. The epigenetic landscape was intended to show that the relationship between genotype and phenotype is non-linear and probabilistic, not deterministic. This importantly incorporates effects of chance or the environment on the eventual outcome.

A newer definition arose with the growth of molecular biology. Here, epigenetics refers to mechanisms of gene regulation that determine the state of a cell and that are heritable through cell divisions but that do not involve changes in DNA sequence. Essentially, this means all the processes that make one cell of an organism different from another, that keep it that way and that allow that state to be passed on to that cell’s descendants. It is often more specifically used to refer to chemical modifications (such as methylation or acetylation) of DNA or of the histone proteins associated with it in chromatin. These epigenetic marks can affect gene expression and can be stably inherited from one cell to another (i.e., through mitotic cell division).

This molecular biology definition has really only a loose relationship to Waddington’s usage. It is obviously true that molecular mechanisms of gene regulation effect (as in mediate) the development of an organism. That is what cellular differentiation and coordinated organismal development entail. Genes are turned on, genes are turned off. Epigenetic mechanisms make the profiles of gene expression that define a particular cell type more stable, with different sets of genes held in active or inactive chromatin conformations. These two usages thus relate to very different levels – one refers to the profile of gene expression of individual cell types and the other to the emergence of the phenotype of the organism.

Now, clearly, the phenotype of an organism depends largely (though by no means completely) on the profile of gene expression of its constituent cells. And there are indeed a number of examples where the behavioural phenotype of an organism has been linked to the epigenetic state of particular genes in cells in particular brain regions. Importantly, such mechanisms may provide one means whereby environmental factors or particular experiences can have long-lasting effects on an organism, by changing patterns of gene expression in particular cells in a stable manner.

This has been demonstrated so far mainly in rodents, but in several different instances (reviewed here and here). These include responses to maternal care, to various kinds of stressors, including that caused by early maternal separation and to other experiences, notably drug exposure. In all of these instances, some environmental trigger or experience induces a response in an animal. One aspect of this response is to alter the set point of the system so that its response to subsequent events of the same type is changed (i.e., learning). In some cases, this involves changes in gene expression and epigenetic marks may help make such changes long-lasting.

The examples above include several where pathways have been worked out in detail, which lead from detection of some stimulus to changes in the chromatin state of specific genes, which are involved in setting the responsiveness or gain of the system. (As in the adjacent figure, from Caldji et al., 2011, showing effects on methylation of the glucocorticoid receptor gene). These may well represent important mechanisms of biological memory for regulating reactivity of various brain systems, which thus influence subsequent behaviour in a long-lasting fashion.

Based on these kinds of examples, epigenetics has become quite a buzz-word in the fields of psychiatric and behavioural genetics, as if it provides a general molecular mechanism for all the non-genetic factors that influence an individual’s phenotype.

Twin studies looking at the heritability of psychiatric disorders or behavioural traits show a consistent pattern: monozygotic twins are considerably more similar to each other for these phenotypes than are dizygotic twins, but are usually not completely identical. This demonstrates an effect of shared genes on phenotypic resemblance (i.e., heritability) but also highlights the limits of that effect – even genetically identical individuals are not phenotypically identical. Some other, non-genetic factors must be contributing to the phenotype of an individual and making monozygotic twins less similar to each other. But does “non-genetic” necessarily mean “epigenetic”?

The fact that environmental factors or extreme experiences can influence an organism’s phenotype is not news. In specific cases like those described above, the effects of such factors may indeed be mediated by molecular epigenetic mechanisms. But here’s the important thing – even though epigenetic mechanisms may be involved in maintaining some stable traits over the lifetime of the animal, they are just that: mechanisms. Not causes. Epigenetics is not a source of variance, it is part of the mechanism whereby certain environmental factors or experiences have their effects. Furthermore, these few examples do not imply that this mechanism is involved in mediating the effects of non-genetic sources of variance more generally.

Differences in the outcome of neural development can and do arise because the cellular events controlling cell migration, axon guidance, synapse formation and other developmental processes are inherently probabilistic. They are determined by the interactions of thousands of different gene products and affected by intrinsic noise at the levels of gene expression and molecular interactions between proteins. The outcome is never the same twice. This is epigenetics in Waddington’s usage – the emergence of a unique organism from a not necessarily unique starting point (the genotype). There is no reason to think epigenetic mechanisms of chromatin regulation are involved in these kinds of differences in neural circuitry.

Note that there are plenty of examples where mutations affecting proteins that mediate or regulate chromatin states (such as MeCP2, CHD7, CHD8 and many others) cause neurodevelopmental disorders such as intellectual disability, Rett syndrome and autism. But these are genetic effects, which disrupt the epigenetic molecular machinery. That is, the important difference between people in these instances is a good, old-fashioned DNA mutation.

So, while epigenetic mechanisms may indeed play a role in the stable expression of certain behavioural tendencies (at least in rodents), it remains unclear how general this phenomenon is. In any case, there is no reason to think of “epigenetics” as a source or cause of phenotypic variance at the level of the organism. And here is a plea: if you are tempted to use the term epigenetic, make it clear which meaning you intend. If you simply mean non-genetic, there is a more precise term for this: non-genetic.

In part 2, I consider a more egregious trend emerging in the literature of late – the idea that transgenerational epigenetic inheritance can provide a mechanism of heredity that explains the so-called “missing heritability” of psychiatric disorders. (It can’t).