Genes and Behaviour
November 30, 2006
Every now and then, the media trumpets the discovery of a gene that is responsible for a particular behavioural trait, creating an impression in the general public that our behaviour is determined biologically. But how valid is this view?
I have remained interested in the issue of genes and behaviour for some time, because I find myself particularly curious about the gaping holes in our body of scientific knowledge, and the tendency for those with a scientistic worldview to ignore or overlook these holes when doing so serves their purposes. (Scientism, incidentally, is the ideology that science has primacy over other interpretations of life, that is, that scientific knowledge is more important than other forms of thought).
A brief history of the gene is required. In the 1860’s Gregor Mendel experimented with cross-breeding pea plants, and came up with principles of inheritance that are still taught today. He hypothesised, based on his observations, that there was a factor that transferred traits from parents to offspring. In 1909, Wilhelm Johannsen coined the term gene for this factor. During discussion in this era of science, all manner of properties were considered to be inherited – including physical traits, behavioural traits, and derived properties such as intelligence and even criminal tendency.
In 1953, Hershey and Chase demonstrated that it was DNA that contained genetic information (and not proteins, as was previously believed), and shortly after several different groups (including Watson and Crick) discovered the helical structure of DNA. There was much excitement in the scientific community, as it seemed that we had learned everything there was to know about inheritance – traits were encoded in genes in the DNA molecule, and this was inherited by the offspring.
However, something rather vital was missed out during this process of the development of ideas of genetics, specifically, it was assumed that everything that was discussed as an inherited trait prior to the discovery of DNA as the means of transmitting genetic information, was indeed transmitted via DNA. In the case of behaviour, this view persisted even in the total absence of any supporting data.
Genetics is a complex subject, but we only need to understand a few simple aspects to conduct this discussion. A gene is a collection of nucleotides that codes for one of two different things: either it is the blueprint for a protein, or it is produces RNA molecules that regulate genes (turn them on or off) or otherwise affect the production of proteins. (Modern research has complicated this view by showing that a gene can code for different proteins, and the DNA for a protein need not come from consecutive sequences, but these points are irrelevant to this discussion).
Let me reiterate this point, as it is crucial: a gene codes for a protein, or it affects the production of proteins. There is nothing else we currently know of that a gene does.
Why, then, is there such talk of a genetic basis of behaviour? Because if behaviour is rooted in genetics, either we are saying that a protein produces behaviour (which we shall see shortly is absurd), or we are claiming that behaviour emerges somehow from genes in a manner we don’t understand and cannot currently prove – at which point the honest scientific position is to admit that we do not know how behaviour is inherited.
Let us look at the claim that a protein produces behaviour by examining a specific case. In 1993, a well-publicised report was published in the journal Science recounting the work of Dean Hamer which claimed to have found a gene which correlated with homosexuality. The media trumpeted this research as having found “the gay gene”. This conclusion was later shown to be flawed, in particular, a study of homosexuality in identical twins demonstrated that homosexuality was not expressed in both twins. Furthermore, studies demonstrated that adoptive brothers show greater incidence of ‘shared’ homosexuality than non-twin biological brothers. This is extremely strong evidence that homosexuality is not genetic in basis.
And we should not be surprised by this, because if there were “a gay gene”, it would mean that there was a gay protein, and that the gay protein caused homosexuality. (This leads to the amusing notion of having a “gay protein shake” – would that make someone gay?) How can a single protein cause a complex behaviour? We have no model that allows for this, and indeed, what little we know of the biological basis for behaviour makes this an untenable claim. It makes for good media sound bites, but it makes for very poor science.
What is the evidence that behaviour is genetic? There is none, although there is some evidence that behaviour is inherited.
The strongest evidence relates to medical disorders. It is certainly the case that problems with specific genes (bugs in the genetic code, if you will) can cause alterations in behaviour. This is not dissimilar to observations that physical damage to the brain can change behaviour. But of course, what we mean when we talk of behaviour in the context of brain damage is not what we always mean when we talk of behaviour – when you try to chat up a potential romantic partner we can call that behaviour, but we are not using ‘behaviour’ in the same context as when we talk about the behaviour that results from damage to the brain.
There have been experiments that demonstrate that altering genes in mice can screw around with mouse behaviour, but this seems to be in the same kind of category – behaviour in the sense of a medical disorder is not the same as behaviour in general terms.
There is evidence that ‘behaviour breeds true’, that is that it is possible for behaviour to be inherited. For instance, consider the behaviour of specific dog breeds, such as the retrieval instinct of a Labrador. But this is evidence for the inheritance of behaviour, not for the genetic basis of behaviour. We can leap the burden of proof if we wish and jump straight to DNA as the basis of this inheritance (and it may not be wholly unreasonable to do so) – but it is not strictly scientific to do so. At best, the idea that all behaviour has a genetic basis is a hypothesis.
Finally, there is the relationship between species and behaviour – in that different species have different behaviours associated with them. But this cannot be considered evidence for a genetic basis for behaviour, since some behaviours alter in species in single generation steps (migration behaviour in birds, for instance) – which would be impossible if behaviour was determined by genes.
Where does this leave us on the subject of genes and behaviour?
The bottom line is that while virtually all behaviour can be influenced by genes, there is no evidence that behaviour is determined by genes. There is no gay gene, no gene for intelligence, no gene for violence, no gene for reliability, no gene for amiability… there is no gene for any behaviour, neither does it seem likely that any such gene will be found. Errors in genes can cause specific medical conditions (which have behaviours associated with them), as with Down’s syndrome for instance, but that is as far as the body of research currently goes.
In 1994, the journal Science published an article by Charles Mann entitles Genes and Behaviour which contained this apposite quote:
Time and time again, scientists have claimed that particular genes or chromosomal regions are associated with behavioral traits, only to withdraw their findings when they were not replicated. "Unfortunately," says Yale's [Dr. Joel] Gelernter, "it's hard to come up with many findings linking specific genes to complex human behaviors that have been replicated. "...All were announced with great fanfare; all were greeted unskeptically in the popular press; all are now in disrepute.
We are nowhere near to understanding the basis for behaviour, although we can say with confidence that complex behaviours do not result from single genes. Because the genes we inherit come in part from one parent, and in part from the other, attempts to tie behaviours to multiple genes seem equally flawed. Something else is going on, we don’t know what, and all we can really do is wait and see what future research turns up on the utterly mysterious subject of behaviour.
This topic is very nice, it shows the different approach from the logically thinking person...:)
Just a small note to the facts contained - DNA never serves as direct matrix for protein replication, information from DNA is transferred to RNA, which is matrix for protein replication... But this is pointless for this discussion:)
Basicaly, I guess that there are proteins which affect behaviour - consider a regular menses cycle:) - this is clear evidence that proteins/steroids hormones can affect the behaviour:). Of course also other hormones affect male and female behaviour - increased testosterone, progesterone level may and will affect your behaviour the same way. But this is not saying anything about the true reason of behaviour and what it has to do with genes.
I suppose that genes doesn't transfer direct behaviour information (with the exception of hormonal information, but this is only altering normal behaviour, not creating it from scratch). The genes just code the information for replication of the organism and its correct functions - something like BIOS, processor instructions sets in our computers we work with.
But this thinking reduces our physical existence to being a "machine", which is the point I don't like.;)
The question about relation between genes and human behaviour directly implies the longest question of religion and god's existence - the answer to which we will only get once we will leave this world:)
The truth is that children take a pattern of the behaviour from parents, but this is not related to genes, but more to the environment children grow in...
Posted by: Slnecnica | November 30, 2006 at 06:41 PM
Disclaimer: I've recently worked with a genetic epidemiology team at the University of Manchester to produce a proof-of-concept of a system to analyse sets of genetic markers for correlation with conditions. It produced results that appear statistically significant. My view may be biased towards this approach as a result.
"How can a single protein cause a complex behaviour?"
You state yourself further down the piece that "It is certainly the case that problems with specific genes [...] can cause alterations in behaviour." You also note that "a gene codes for a protein, or it affects the production of proteins." It would strike me, therefore, that combining these two statements, you might accept that a gene could affect the production of one or more proteins, and that together those alterations in proteins could cause other problems - for example a reduction in serotonin production? (I'm not saying you *do* accept the above; I wouldn't put words in your mouth. I'm asking).
Much of the scientific research at the moment is focused not on single genes (or on single proteins) but on pathways: the set of processes that are required to manufacture a specific product in a specific quantity within a specific cell. For some of these processes, it's been demonstrated that interfering with one gene alters the process (I'll hunt down some of the papers if you're interested); for others, nobody presently has a foggy clue what the mechanism is.
For an interesting third set, it's possible to demonstrate a statistically very significant (1 in 10^6 or lower odds of it being chance using analysis techniques presently in favour in the statistical community - again, references on request). This is where the project referred to above comes in. The state of the art is now at the stage where it's possible to get large numbers of markers (not the same as sequencing the DNA - this gives a mere 10^5 to 10^6 markers per individual at present) from large numbers of individuals relatively cheaply. Cohort studies are also at the stage that it's possible to find cases and controls for many common conditions (rheumatoid arthritis and bipolar depression are two being investigated at the moment). So we're starting to get to the point of being able to crunch large datasets and find the correlations, and then being able to (probably) exclude the inevitable false positive results of the fishing expedition - if you look through enough random data, you'll find some correlations purely by chance, and you need to get rid of as many of these as possible, preferably without getting many false negatives. The study for the proof-of-concept was using 500k markers across 5k individuals, and we were looking for the harder-to-find signals: pathways consisting of more than one stage, controlled by multiple markers. There's a large compute cluster in Manchester chewing on this problem as I write.
Why is this relevant? Because bipolar depression is a behavioural condition. I don't know what you'd consider "proof" - and you know I follow Popper on *that* subject - but if this system throws up a very strong correlation between particular markers and the incidence of bipolar, I would think that would at least warrant further exploration as a pathway that might have a causal link with the condition, might cause a predisposition, or might aggravate other factors in combination. Yes, this is a much weaker statement than "a gene for"; but also a much more precise one. I'm as peeved as you are when The Meeja get out their trumpets and distort a measured specific statement into a black and white generality!
"there is no gene for any behaviour, neither does it seem likely that any such gene will be found."
Are we being a little strident, Chris?
I assume you meant something like "there is no proof at present of a causal link between gene and small variations in behaviour", as the statement you made asserts that you believe that no such thing exists across all species, all time and all behaviours however gross, and is apparently contradicted by your statement about gross variations in mouse behaviour?
Can you perhaps tell us a little more about the way in which you use the word "behaviour" in this post? You're apparently excluding gross changes in behaviour from consideration, and I'm somewhat concerned that you're setting the bar as high as the detractors of "artificial intelligence", where any problem we know how to solve on a computer, such as playing chess, is immediately removed from the domain of AI by the folks who view AI as "impossible". What are the terms of reference here?
Posted by: Peter Crowther | November 30, 2006 at 08:42 PM
"How can a single protein cause a complex behaviour?"
How can pressing a single button on a pinball machine cause such complex activity to result? A cell contains a rich soup of proteins (amongst other things), with many causal chains and possibly some loops in there. Consider being able to construct a directed graph (possibly acyclic, possibly not, but definitely not a tree) of causal links of a protein affecting the generation or expression of other proteins. Now change the level of one protein in that graph. What happens from there? I suspect it involves a number of other proteins, and probably some quite sophisticated negative feedback systems in any real cell to ensure small perturbations do not destroy the cell.
Cells are not simple devices, and treating them simplistically would not seem to be appropriate.
Posted by: Peter Crowther | November 30, 2006 at 09:40 PM
Slnecnica: Of course, to keep this piece brief I left out a lot of details. ;)
"But this thinking reduces our physical existence to being a "machine", which is the point I don't like.;)"
I sympathise! The media has used genetics to present something akin to this view, and this is largely my point of objection.
Peter: One topic that I left out completely is the issue of the brain and behaviour, because this is so vast an area that I couldn't hope to cover it here. But with the absence of this topic, it makes it difficult to cover the issues with any completeness. Perhaps we would do better to return to this at a later date when this has been added to the table.
However, in the interest of doing your comment justice for the time being...
The important thing to remember is that when I talk about 'behaviour' here, I mean the full range of animal behaviour - hunting, mating, shopping, politics, art, cooking, play, love, war etc. Conditions such as depression, bipolar disorders etc. although behavioural fall under my comment that "behaviour in the sense of a medical disorder is not the same as behaviour in general terms".
Behaviour can be *affected* by single proteins and hence single genes, but behaviour is not *determined* by genes.
The point being made here is that "free will" (which I did not name check because of the can of worms that would wrench open) is not refuted by anything we know about genetics. In fact, since we do not really know how complex behaviour comes about, all the cards are still in the air at this time.
Regarding your claim that complex activity results from pressing a button on a pinball table - the behaviour of a pinball table is comparatively very simple, considerably less complex than a cell, let alone an organism, let alone the behaviour of an organism. Furthermore, our current scientific models explain all the mechanisms in the operation of a pinball table. They do not explain all the mechanisms in genetics, cell operation, cell co-ordination, brain construction, brain operation and so forth.
Cells are largely irrelevant to behaviour as I am talking about here, just as atoms are irrelevant to marketing and quarks are irrelevant to painting (generally speaking).
Genes certainly affect the operations of cells. But cells do not *by themselves* determine behaviour. Behaviour is a complex emergent property of biological life, the causes and roots of which are not at all understood. We have observations of behaviour, and we have knowledge of medical conditions that disrupt behaviour in predictable ways. But we do not have anything that links the observations we have of behaviour to anything tangible in the biological substructure.
Your position seems to boil down to: since larger phenomena can be affected by smaller phenomena, how do you rule out behaviour being affected by single genes?
But I do not rule out behaviour being *affected* by single proteins/genes, I rule out behaviour being *determined* by single proteins/genes. Or indeed *resulting* from single proteins/genes.
Perhaps I did not cite sufficient examples of what I mean by behaviour (that is, perhaps I took for granted what behaviour means in the wider sense)...
Here's an attempt to expand the question into (randomly chosen) specific terms:
How can a single gene result in cuckoos acting as brood parasites?
How can a single gene result in social hierarchies among wolves?
How can a single gene result in a sexual preference in humans?
It cannot. A single gene can be *involved* in these complex behaviours, but it cannot cause or determine them by itself - or at least, such is my opinion based on what I have studied.
However, a single gene can apparently *disrupt* complex behaviours in predictable ways - but those complex behaviours must first exist in order to be disrupted. A gene might be linked to depression, say, but the experience and behaviour of depression results from the state of being a sentient organism, which emerges from complex biological mechanisms far beyond the scope of a single gene - and also beyond our current understanding.
And that is the point I am trying to make here. Because the popular media advances a view of genetics that suggests to many people, less educated than yourself, that genes *determine* the full range of behaviour - that it is meaningful to talk of a gene for shoplifting, or skill at poetry, or liking the colour blue - and it is this misconception that I was attempting to challenge in this piece.
Hope I have made my position clearer!
Posted by: Chris | December 01, 2006 at 10:27 AM
Chris - many thanks for the clarification. I think we agree, with one possible exception: I don't hold that any behaviour is necessarily "normal", and therefore describing "disruption" to "normal" behaviour feels somewhat odd to me. My own position is closer to there being alternative behaviours that are equals rather than a "normal" and one or more "disrupted" forms. Given that difference in terminology, I suspect our apparent difference (that you see a single gene as being able to disrupt normal behaviour, whereas I see it as being able to affect behaviour by flipping the behaviour between two equals) is merely apparent. However, we may have a deeper disagreement on whether one behaviour of a set can be characterised as "normal" (rather than, say, "most common in the observed population")!
Posted by: Peter Crowther | December 01, 2006 at 12:59 PM
Yes, I think we do broadly agree. Even I am not convinced by my own wording in terms of "normal behaviour" - but I think we both know what I am trying to say, so perhaps in this instance we can save me having to find exactly the terminology I'm looking for. ;)
PS: I got (and hid) your earlier comment - will try and give you a call when I get back from Slovakia next week. Best wishes!
Posted by: Chris | December 01, 2006 at 01:05 PM
Posted by: Anon | March 07, 2007 at 12:31 AM
Very interesting reading indeed. Some weeks ago I had a discussion with three friends precisely on this. I had a hard time trying to make them understand what a gene was (there are different definitions) and so I stuck to the conventional definition: genes code for protiens,which is a very different statement than a gene for homosexuality (interesting that nobody, as far as I know, has found a gene for heterosexuality or bisexuality-correct me if I'm wrong). While a gene (or better,a number of genes working in tandem )may predispose an individual to behave in certain ways, the role of the environment is crucial,inasmuch as it may trigger the gene(s) in the first place. And here lay my problem: In Dawkin's terms it is very difficult to determine, with precision, the nature/nurture aspect of human behaviour. Dawkins uses the analogy of baking a cake whereby it is impossible to unbake a cake! It may be also profitable to cast the debate as nature and nurture working together, rather than against each other. The complexity of life shows that cellular activity cannot be simply multiplied on the level of an individual. I guess, what I'm trying to say is that complexity itself is a daunting issue. Unfortunately, the last three years have been taken up by non-scientific studies and I'm definetely not up-to-date with current research regarding genes and behaviour. And yes, I get so very irritated when a newspaper claims that a scientist, or a group of them, discovered a gene for IQ or a gene for aggressiveness. The report usually takes the form of a short article or a paragraph summarising complex data. There is no justice done to the way that serious scientists communicate their findings, ie, the scientific paper.
Cheers and well done. I wouldn't mind feedback on my comment.
Posted by: Nicholas Zarb | August 06, 2007 at 12:11 AM
Nicholas: thank you for your comment! Like you, I find myself frustrated by what is overstated in respect of genes. Modern biology has come a long way - yet there is still so much still do be done.
The effect of environment on biology is exceptionally difficult to disentangle - simple factors, such as temperature and sunlight can affect biological development (look at different plants when grown from related seeds but in different altitudes and climates). Then, for humans, we have the effect of culture - part of what constitutes our behaviour does not appear to be biologically determined in any way.
The failure to find "a gene for heterosexuality" is a sign that the search for a "gay gene" *presupposed* that being gay was an anomalous condition. :)
I completely agree with you in regards to the complexity of what is being asked here - we like to think of science as being on top of its knowledge, but in fact science only "knows what it currently knows", and the science of the future rarely resembles the science of the past.
Who knows to what extent future generations will laugh at our ignorance! :D
Take care, and thanks for commenting!
Posted by: Chris | August 06, 2007 at 04:34 PM
While it seems highly misleading to talk about a single gene for something as complicated as, say, sexual preference, it seems obvious that a collection of genes does precisely that.
I'm talking, of course, about the X and Y chromosomes, whereby a single pair of chromosomes split the population in half, and instill the majority of the population with a strong preference for the other gender.
There are a lot of other effects as well, but as a collection of genes, a chromosome is limited in it's effects to a collection of proteins, and the result is quite clear.
While my own uninformed assumption is that the search for a single "gay gene" will be fruitless, I would be very surprised if there was not a complex interrelated group of genes - possible several different groupings - that totalled a preference for the same gender in the same way that the majority of people are attracted to the opposite gender.
(The fact that homosexuality is expressed in other species strongly suggests that it is not cultural influences at work, but does not rule out different causes for the same effect).
Posted by: Jules | October 29, 2007 at 06:38 AM
Jules: there are of course genuine consequences of genes, but usually structural - as in the case of male and female gender. I am doubtful that a cluster of genes will explain sexuality; I rather suspect it has something to do with the process of imprinting, a mental process that is quite possibly independent of genetics (except in so much as genetics contributes to the contruction of the brain, I suppose).
As you say, homosexuality occurs in most mammal species, so dismissing it as cultural isn't an option. But also as you say, this does not rule out non-genetic causes. I'm betting on imprinting, but I don't have data to support my position.
Thanks for the comment!
Posted by: Chris | October 30, 2007 at 12:35 PM