Warrior genes and the disease of being a scientist

By Hilary Miller 15/09/2009

The past few days, headlines like ’Maori don’t have warrior gene’ and ’Maori warrior gene debunked’ have been all over the media. This has left me with a sinking feeling in the pit of my stomach, and thinking that this sounds a lot like media hype/oversimplification of what is a very complex area of research.  To recap…

Back in 2006, Rod Lea gave a presentation at the 11th International Congress of Human Genetics showing that Maori have a higher frequency of a particular variant of the Monoamine Oxidase-A (MAO-A) gene.  In some studies, this particular variant has been linked with aggression and antisocial behaviour, and one study back in 2004 dubbed it ’the warrior gene’.  The media picked this story up, and bandied around headlines like ’Warrior gene blamed for Maori violence’, making statements claiming that “New Zealand Maori carry a “warriorgene which makes them more prone to violence, criminal acts and risky behaviour’.  This is not what Lea and colleagues claim in their original study at all — I’ll talk more about that below.

Anyway, now according to media reports this claim has been “debunked by science”.  When I read this my initial thought was that someone has done another study of Maori MAO-A allele frequencies, and found conflicting results.  But actually this is not the case at all.  The ’scientific study’ that debunks this claim is actually just a review by Maori academic Dr Gary Hook, published in Mai Review — a peer-reviewed journal of Maori and Indigenous development, but not a scientific journal.  Hook makes some good points, which I’ll talk more about in a minute, but presents no new data and much of his review of the scientific controversy has already been covered in a previous article.

So what is monoamine oxidase, and what did Lea and colleagues actually find in their study? Monoamine oxidase enzymes break down neurotransmitters like serotonin and dopamine, and are therefore capable of affecting mood.  These proteins and the genes that code for them come in two forms — A and B — it is the A form that is the subject of their study.  This gene contains a number of variants, one of which contains a 30 bp repeat (MAO-A30bp-rpt) in the promoter region of the gene.  The number of times this 30bp sequence is repeated affects how active the gene is and therefore how much MAO-A enzyme is produced.  The 3-repeat form (or allele) in particular results a lower level of MAO-A activity and higher dopamine levels.  Several genetic association studies have linked this low activity variant with antisocial behaviour and increased aggression, but the results are rather complex.  The three largest studies have all found that there is actually no relationship between this variant and antisocial behaviour when this is analysed in isolation, but that this allele is associated with behaviour when environmental factors are taken into account.  For example, a study of Dunedin children found that those who who were abused and neglected in childhood were more likely to show antisocial behaviour later in life if they had the low activity form.  It is important to note that the studies that found this effect were all on caucasian males, and studies of other ethnic groups found no such effect. 

In response to the media storm created by their intial findings, Rod Lea and Geoff Chambers presented a brief outline of their work in the New Zealand Medical Journal in 2007.  I haven’t seen the full study published in a peer reviewed journal, but I suspect it may be in the works.  Anyway, their report from 2007 states that they genotyped 46 Maori males, and found that the low activity allele was present at a frequency of 56% – significantly higher than the frequency found in caucasian males.  They also claim evidence for positive selection on the MAO-A gene, implying that the low activity variant conferred an advantage at some point in Maori evolutionary history.  They suggest that this variant “may have conferred some selective advantage during the canoe voyages and inter-tribal wars that occurred during the Polynesian migrations”.  Their evidence for selection comes from a much smaller sample size of only 17 individuals (as individuals without 8 Maori great-grandparents were excluded to reduce the effect of European genes) and it is hard to evaluate from the information given in the NZMJ article.  It is possible that a larger sample would show different results, or that a genetic bottleneck associated with the colonisation of NZ would result in the same pattern as a result of chance sampling of alleles.

The work on MAO-A by Lea and colleagues was part of a study aimed at analysing the MAO-A gene as a genetic marker for alcohol and tobacco response traits in New Zealanders.  An important part of these types of studies is identifying whether there is any ethnic variation in MAO-A allele frequencies that can confound the results.  Thus their motivation was not to find a “gene” for aggression in Maori, and at no point do they claim that high frequencies of the MAO-A low activity variant cause increased violence and antisocial behaviour in Maori.  This is not what they tested, and in fact no other study has tested this either.  In their NZMJ article, they state:

It is important that the incidental formation of this ’warrior gene hypothesis’ is interpreted for what it is–a retrospective, yet scientifically plausible explanation of the evolutionary forces that have shaped the unique MAO-A gene patterns that our empirical data are indicating for the Māori population.
As alluded to by Merriman and Cameron, the extrapolation and negative twisting of this notion by journalists or politicians to try and explain non-medical antisocial issues like criminality need to be recognised as having no scientific support whatsoever and should be ignored.

The way Lea and Colleagues framed their study did not help their cause though.  They stated that the MAO-A variant has been strongly associated with aggression and risk taking, without adding the qualifiers that this is only in some ethnic groups, and only when previous environmental factors are taken into account.  This, in addition to referring to the MAO-A variant as the ’warrior gene’, was bound to get misinterpreted by the media. 

Gary Hook’s article is largely concerned with the cultural issues around branding Maori with a ’warrior gene’.  This is an opinion article, not a scientific study, and much of the scientific “debunking” of the warrior gene hypothesis is simply a review of conclusions already presented by Merriman and Cameron in the NZMJ.  Hook’s main point (which I think is good one) is that labelling maori with a “warrior gene” is akin to labelling them with a disease, and that propagating the idea that maori are intrinsically violent is dangerous and unhelpful.  He rightly states

the implications that follow from the “warrior” gene hypothesis should it become fact in the minds of the general public are horrendous

The media attention given to this article will go a long way towards ensuring that this hypothesis does not become fact in the minds of the general public, so in that sense it has served its purpose.

I think Hook does science a disservice though.  In the introduction he states

 It was proposed that the high criminality of Māori was due to the expression of a “warrior” gene that rendered Māori “more prone to violence, criminal acts, and risky behaviour.”

citing both a news article and Lea and Chambers’ NZMJ article.  The media may have claimed this, but Lea and Chamber’s article certainly didn’t.   He also unfairly states

It is one thing for newspapers to promote their fetishes but it is another for scientists to be the source of speculation and fantasy about the nature of Māori

 when the speculation and fantasy came from the media, not the scientists. 

The associated media coverage of this report does nothing for understanding the science, and contributes to the “scientists always get things wrong” attitude that seems to be prevalent out there.   To present the science in this case as a series of black and white facts – first we think they have a warrior gene, and now we think they don’t – is extremely misleading and only adds to the misunderstandings surrounding the original study.


 Hook, G.R. (2009). “Warrior genes” and the disease of being Maori. Mai Review, 2, Target article.

 Lea, R., & Chambers, G. (2007). Monoamine oxidase, addiction, and the “warrior” gene hypothesis. Journal of the New Zealand Medical Association 120(1250)

 Merriman, T., & Cameron, V. (2007). Risk-taking: behind the warrior gene story. Journal of the New Zealand Medical Association 120(1250)