Finnish ‘violent gene’ study – experts respond

Two genes that may be associated with violent behaviour have been identified in a study of Finnish criminals, but independent experts are urging caution in interpreting the results.

DNA genomicsA new study analysing the DNA of 895 Finnish individuals guilty of criminal behaviour found a link between criminal behaviour and variations in two particular genes: cadherin 13 (CDH13) – a gene involved in neural connectivity – and monoamine oxidase A (MAOA) which is involved in controlling levels of dopamine in the brain. The MAOA variant, also dubbed the ‘warrior gene’, was controversially linked with Maori populations in 2006.

In the conclusions of their article, published in Molecular Psychiatry, the authors state, “a conservative estimate implies that 5-10% of all severe violent crime in Finland is attributable to specific MAOA and CDH13 genotypes.”

Our colleagues at the UK SMC collected the following expert commentary. Feel free to use these quotes in your reporting. If you would like to contact a New Zealand expert, please contact the SMC (04 499 5476;

Dr Seena Fazel, Wellcome Trust Senior Research Fellow, Department of Psychiatry, University of Oxford, and Honorary Consultant, said:

“This is a potentially interesting study that needs replication in larger cohorts and should also be interpreted cautiously as recent reviews of MAOA genes have not found clear associations with violence.

“The field would benefit from larger sample sizes to improve precision combined with additional approaches, including the investigation of gene-environment interactions and the role of epigenetic regulation in aggression. Working out the most informative markers is also important – currently the cadherin 13 (CDH13) marker is potentially interesting and needs further replications. It is really key though that we have a better characterisation of what we mean by ‘severe violence’ as we all need to be talking about the same thing.”

Prof John Stein, Emeritus Professor of Physiology, University of Oxford, said:

“This is a very interesting study with plausible aspects. But please do not accept the claims that these alleles are ‘responsible for 5-10% of violent offences in Finland’.  All they show is that they may contribute 5-10% to the chance of an individual being very violent. In fact their Genome Wide Association Study (GWAS) did not give any significant results. So, if a single genetic factor alone could significantly explain violent behaviour they would have seen it in their GWAS results. These alleles are quite common and so environmental factors are probably much more important.  For instance simply improving prisoners diets can reduce their violent offending by 37%.”

Prof Jan Schnupp, Professor of Neuroscience, University of Oxford, said:

“This study is not particularly new. Studies describing an increased incidence of genes coding for low activity monoamine oxidase (MAOA) enzymes among violent offenders have been around for a long time.

“When interpreting these results, it is very, very important to put them in the right perspective – which means not to get excited about them. At all.

“The thing that should be mentioned again and again is that HALF THE POPULATION have low-activity MAOA enzymes*, but only a tiny fraction of these carriers end up behind bars for violent crimes.  The overwhelming majority of carriers of these genes, >99.5% or so, manage to make it through each day without giving in to impulses to bash other people’s heads in. Half the people in your office will carry these genes. Odds are 50/50 that you do. How violent has your day been? To call these alleles “genes for violence” would therefore be a massive exaggeration. In combination with many other factors these genes may make it a little harder for you to control violent urges, but they most emphatically do not predetermine you for a life of crime.

*) look at fig 1 of the paper. The y-axis is an odds ratio of carrying the allele. For people without violent crimes, the odds ratio of carrying the “violence gene” is about 1, i.e. it’s 50/50! For people with 10 crimes it edges up to 2 or 2.5, meaning the odds become 70/30 or thereabouts. It is a classic case of an effect that has strong statistical significance but only a very small effect size.

Dr Malcolm von Schantz, Reader in Molecular Neuroscience, University of Surrey, said:

“Behavioural genetics is a very interesting area of research, but also one that is full of controversies. There is both the issue of whether a genetic association can be replicated, and the issue of how to interpret it – the public will be asking themselves if the scientists suggesting that violent offenders should not be fully accountable for their actions.

“This publication is interesting on a number of levels. Finland really is in many ways the promised land of human genetics, with a population that has been relatively isolated for a long period. So the Finnish people share a relatively small number of common ancestors, with a large amount of interbreeding and a limited addition of foreign elements to the gene pool. The other thing which is so unique about Finland is their very organised record keeping, in terms of health data and, in this case, the National Prison Register. Thirdly, Finnish people are generally very generous in agreeing to participate in genetic studies – obviously this study, like any other, can only be performed if people consent to take part in it.

“This study combines the old candidate gene methodology, where you look at the function of known genes and make hypothesis about how they might relate to a certain disease or behaviour, and the newer genome-wide association study methodology, where you look at millions of markers across the entire genome without any prejudice for or against any gene.

“The candidate gene where an association has been found, MAOA, relates to two previous famous genetic association studies of violent offenders which provided a credible biological mechanism – a reduced function in an enzyme which is a target for may drugs used to treat depression and anxiety. However, there has been a big question mark remaining around these studies because of the lack of replication. This reports presents us with yet another piece of evidence that there is likely to be something to the MAOA story.

“The other exciting part of this paper is that the GWAS study yielded a novel candidate CDH13, which produces a neuronal membrane adhesion protein. Even though the GWAS study is done without any preconceptions about function, the gene that came out has already been reported to associate with a wide range of mental health problems. And the authors were able to replicate this hit in an independent population sample.

“So does this paper bring us closer to a situation where violent criminals can claim diminished responsibility because of the genes that they were born with? I think we have to remember that it becoming increasingly clear that there is not one single genetic variant that has a large effect on this, or indeed any complex behaviour. The pattern that is emerging is one of many genetic factors where each one has a small predisposing effect. “The authors themselves calculate that 5-10% of violent crime in Finland is attributable to these genetic differences. I don’t think that means that 5-10% of crimes are individually 100% attributable to these genes. Rather, it would mean that they could have had an influence of 5-10% in each case. So it is not a case of replacing free will and criminal responsibility with a genetic explanation. But I think findings such as these may make it possible, in future, to screen people with vulnerable backgrounds and identify those who are at greater risk of becoming offenders, so that they can get appropriate help before the commit any serious violent crimes.”

Dr William Davies, Senior Lecturer, Behaviour Genetics Group, MRC Centre for Neuropsychiatric Genetics and Genomics, Cardiff University, said:

“The study by Tiihonen and colleagues is an interesting, and potentially important, one which indicates two genetic regions where variation may influence the likelihood of being convicted of multiple violent crimes. Should these preliminary findings hold up, they would have significant implications for: i) identifying potentially violent offenders at an earlier stage and therefore for implementing suitable early interventional strategies and ii) for identifying neurobiological pathways that might be amenable to treatment

“Crucially, the current findings should not be portrayed as a demonstration of ‘genes for violence’ (notwithstanding the need for additional replication); many individuals possessing one or more copies of the so-called risk variants must not have been violent at all, and conversely, several individuals possessing no risk variants were in the ‘ultra-violent’ group. It is likely that these variants only predispose moderately to impulsive or aggressive traits, which, if harnessed in the correct manner, would not result in deleterious consequences for society. For example, such traits might be beneficial in the context of some sports.

“Their first finding, that one version of the X-linked monoamine oxidase (MAOA) gene is present in ~60% of males who have been convicted of at least 10 violent crimes but only ~40% of non-violent males, apparently substantiates previous controversial suggestions that this version of the gene is in some way associated with violent criminality and highlights a potential mediating neurobiology (metabolism of the monoamine neurotransmitters dopamine and serotonin).

“Their second finding, that a version of the CDH13 gene was present in at least one copy in ~52% of repeatedly violent offenders but only ~32% of non-violent subjects, is brand new, and came from an unbiased screen of the whole genome. This finding is intriguing in that CDH13 it has previously been associated with other disorders of impulsivity, and it encodes a protein with a known role in the brain which may be amenable to drug targeting. It is important to note that this second finding did not quite meet accepted statistical criteria for significance, and will need replication in larger sample sizes. Moreover, it will be important to replicate both genetic findings in similar cohorts from different countries to confirm that they do not simply result from DNA features specific to the Finnish population, and/or from the idiosyncrasies of the Finnish justice system.

“Other issues that could potentially bias this study are: i) the large number of prisoners who refused to participate (~20%), who could feasibly be very different in terms of their experiences and genetics to the consenters, and ii) the fact that rape does not appear to have been classified as a violent crime, but rather a sexual crime (and offenders excluded from subsequent genetic analyses).

“Finally, it is formally possible that these genetic variants are not associated with degree of violent offending per se, but rather with an inability to avoid detection for crimes perpetrated.”