In the second week of August 2006, the media were gripped with the announcement that Māori men were genetically predisposed to “violence, criminal acts, and risky behaviour.” (Christchurch Press, 9 August 2006). The basis for this release was the report of a study that purported to show that a genetic variant of the monoamine oxidase (MAO)-A gene, dubbed the “warrior gene” by the media, had previously been “strongly associated with risk taking and aggressive behaviour” and was “strikingly over-represented” in Māori men.1

Because these findings were released at an international conference by a government-funded researcher, there has generally been an assumption that the link between the MAO-A variant and aggression in Māori is based on robust scientific evidence. Here we place the scientific evidence under closer scrutiny.

The regulatory region of the human MAO-A gene has a genetic variation consisting of a small stretch of DNA repeated a variable number of times.2 The 3-repeat form of this ‘variable number of tandem repeats (VTNR)’ polymorphism (present on 33% of Caucasian chromosomes) exhibits up to 10-fold less activity than the 4-repeat form (present in 65% of Caucasian), and has been commonly termed the ‘low-activity’ variant.2

The MAO-A gene controls the production of the MAO-A enzyme, which is involved in the breakdown of several neurotransmitters in the brain, such as dopamine and serotonin. It is believed that this enzyme is important in preventing the build-up of an excess of these transmitters. In MAO-A deficient mice that lack the enzyme altogether, increased levels of neurotransmitters were documented and the mice were observed to exhibit aggressive behaviour.3

In a Dutch family, a dysfunctional MAO-A gene has been linked to antisocial behaviour and increased aggression amongst humans.4 Based on this observation, the MAO-A VNTR has been tested for association with antisocial behaviour in population-based male cohorts. The three largest studies (all in groups of Caucasian ethnicity) published remarkably consistent findings; no main effect was found for the relationship between the MAO-A VNTR genotype and antisocial behaviour, but the high-activity form was associated with buffering males who were abused and neglected in childhood from increased risk of anti-social behaviour later in life.5–7 That is, there was evidence for a gene-environment interaction where the effect of the MAO-A low-activity variant was dependent on the environment—analysed in isolation the gene variant could predict nothing about aggression in male carriers.

This effect was, however, not evident in a cohort of American non-whites.7 A possible biological explanation of the genetic findings is that the higher-activity variant produces more MAO-A enzyme and is better able to mop up excess neurotransmitters, such as dopamine and serotonin, in the brain.

It is clear from the genetic data that MAO-A itself is not associated with aggression (contrary to the interpretation by Lea and colleagues of “strong” association with aggressive behaviour1). More generally, it is important to emphasise that the strong, clear, and direct causal relationship implied by the phrase of “a gene for...” does not exist for psychiatric disorders.8

The MAO-A gene was termed the “warrior gene” on the basis of experiments in Rhesus macaque monkeys.9 These primates have a comparable VNTR upstream of the MAO-A gene with the 5- and 6-repeat forms having 1.3-fold greater activity than the longer 7-repeat form.9 However, in contrast to the human gene, in the Rhesus macaque, fewer repeats means more MAO-A enzyme is produced.

When tested for association with aggressive behaviour in 45 male macaque monkeys there was no main effect of genotype on aggression. There was, however, evidence for a gene-environment interaction; in peer-reared monkeys, the higher-activity allele was associated with aggression, whereas in mother-reared monkeys the lower-activity variant was associated with aggression.9

This gene-environment interaction superficially conflicts with the human data, although there are numerous reasons (including species difference) that could account for this apparent conflict.9 Nevertheless, on the basis of one un-replicated experiment on only 45 animals, the MAO-A gene was (dubiously) termed the “warrior gene” at the 2004 Annual Meeting of the American Association of Physical Anthropologists by a scientific journalist.10

As far as we are aware, the term “warrior gene” has not been applied to humans in previous studies of MAO-A and aggression. Dr Lea and colleagues reported that the frequency of the 3-repeat short form of the MAO-A VNTR was just over 60% in a very small sample of 17 Māori males.1 Based on these data they concluded that “positive selection of MAO-A associated with risk taking and aggressive behaviour has occurred during the Polynesian migrations” and termed the low-activity variant the “warrior allele”. The MAO-A gene was then linked to antisocial behaviour in contemporary Māori males in the publicity surrounding the conference presentation.

We believe that this conclusion is based on science with insufficient investigative rigor, both in interpreting and applying the relevant literature, and in generating new data.

Central to the argument of Dr Lea and colleagues1 is the assumption that the low-activity MAO-A allele is associated with aggression in Māori males. This assumption cannot be made without the appropriate genetic epidemiological experiments being done to test for an association between MAO-A and aggression. However no such study has ever been reported.

The lack of evidence for involvement of this variant with violence and antisocial behaviour in American non-whites7 demonstrates that extreme caution is needed when translating MAO-A genetic findings between racial groups. A central tenet of complex phenotype genetics is that genetic associations are likely to vary between racial groups.

There is no direct evidence to support the claim that the MAO-A gene confers “warrior” qualities on Māori males, either modern or ancestral. Furthermore, the assumption that a genetic association in Caucasian applies in Māori; the use of the “warrior gene” label in the context of human MAO-A aggression studies; generalising from a sample of 17 individuals not representative of the general Māori population; and the lack of scientific investigative journalism have combined to do science and Māori a disservice.

Author information: Tony Merriman, Senior Research Fellow, Department of Biochemistry, Otago School of Medical Sciences, University of Otago, Dunedin; Vicky Cameron, Research Associate Professor, Christchurch Cardioendocrine Research Group, Christchurch School of Medicine and Health Sciences (University of Otago), Christchurch

Correspondence: Dr Tony Merriman, Department of Biochemistry, Otago School of Medical Sciences, University of Otago, PO Box 56, Dunedin. Fax: (03) 479 7866; email: tony.merriman@stonebow.otago.ac.nz