Skip to content

When you choose to publish with PLOS, your research makes an impact. Make your work accessible to all, without restrictions, and accelerate scientific discovery with options like preprints and published peer review that make your work more Open.


Comparing Adam Lanza’s DNA to Forensic DNA Databases: A Modest Proposal

Is there a genetic signature for criminality? It’s an old and controversial question. (NHGRI)

In 1729, Jonathan Swift of Gulliver’s Travels fame published a satirical essay called “A Modest Proposal.” He suggested that a cure for poverty was for poor people to sell their children to rich people as food.

I’m borrowing Swift’s title to bring up another outrageous idea: analyzing forensic DNA databases for a genetic signature of criminality.

Days after the Newtown shootings of December 14, 2012, headlines trumpeted the state medical examiner’s request of University of Connecticut geneticists to examine mass murderer Adam Lanza’s DNA. What exactly that might entail wasn’t announced, but celebrity docs, geneticists, and bloggers weighed in, nearly all agreeing that (1) violent tendencies are due to complex interactions of many genetic and environmental factors and (2) probing Lanza’s DNA and finding anything even suggestive of causing his crime could lead to stigmatization of individuals who share suspect genome regions with him.

Behind the denials of a genetic explanation for criminality lies a history of just such associations.

Past candidates for criminal DNA, listed in many articles last week, include the extra Y chromosome of the 1960s and the monoamine oxidase (MOAO) mutation behind a Dutch family of rapists and arsonists, described in 1993. Shortly after, researchers identified a different gene variant that tracked with violence and suicide in Finnish families.

In January 2012, criminologists published a study that applied a “delinquency scale” to assess whether such behaviors as painting graffiti, lying to parents, running away, and stealing, were more likely to affect identical twins than fraternal twins, suggesting a genetic component. The headlines that their article in Criminology spawned, with the help of news release hype, were predictable: “Life of crime is in the genes, study claims.”

If an investigation of petty crimes inspires such strong headlines, the fear of unleashing genetic discrimination from analyzing Lanza’s DNA seems justified. Yet it appeared odd to me that several articles deemed any response to the sequencing of the killer’s DNA unlikely, because it would be a sample size of one.

We do, in fact, have sources of criminals’ DNA. And they’re extensive.

A blog from the Council for Responsible Genetics, for example, claims that “Focusing on the results of the study [on Lanza] could also prove problematic since there is (sic) basically no data to compare it to,” then quotes a University of Massachusetts Medical School professor saying “we don’t have enough of a sample size.”

But forensic DNA databases in many countries have been storing the DNA of convicted criminals since the mid-1990s, many killers among them.

The UK led the way in DNA profiling (I wrote the cover story on it for Discover in June 1988), and their National DNA Database now has samples from more than 6 million individuals.

In the US, the Combined DNA Index System (CODIS) has more than 10 million samples. Thailand just signed on to use CODIS on voluntary samples from 100,000 inmates, and 39 other nations already use the system.

The lines of a DNA profile represent the numbers of short repeated sequences at 13 sites in the genome. (NHGRI)

CODIS generates a DNA profile for an individual based on 13 sites in the genome that vary in the number of repeats that they harbor. One such “short tandem repeat” (STR), for example, includes the DNA sequence “GATA” present in 5-16 copies on each of a person’s two chromosome 7’s. For that marker alone, 78 combinations are possible. Multiplying the frequencies of the different variant (allele) possibilities in a particular population for all 13 markers generates enough diversity to distinguish individuals.

Within the STR DNA profiles of these millions of convicted individuals may emerge a genetic pattern that’s more common among mass killers like Lanza. Maybe significantly so. And if researchers have access to DNA samples and not just CODIS profiles, they could, theoretically, compare any part of the genome. If there is such a thing as measurable inherited criminality, then as the numbers build in the databases, associations between DNA patterns and certain behaviors may strengthen, perhaps even suggesting a mechanism that can be used in drug discovery or repurposing. (I readily admit to not knowing the legalities of using forensic data for new purposes; I’m hoping an attorney will weigh in. I’m just the gene girl.)

A very large control group would also be necessary to weed out potential false positives, like showing that a disease-causing mutation is found only among patients.

DNA forensic data could and should be de-identified, because the crimes are important, not the names. According to the FBI CODIS fact sheet, “If all personally identifiable information is removed, DNA profile information may be accessed by criminal justice agencies for a population statistics database, for identification research and protocol development purposes, or for quality control purposes.“ And informed consent isn’t required of convicts.

Would use of a genetic signature for criminality plunge us into the world of Minority Report, the 2002 Tom Cruise film in which police in a dystopian society arrest people before they’ve committed crimes? I would hope not. But I can imagine a scenario in which a psychiatrist uses such a genetic test for a patient whose background suggests violent tendencies. The patient wouldn’t suffer Tom Cruise’s fate of premature punishment, but perhaps wouldn’t be allowed to purchase a gun.

A powerful argument against the use of forensic DNA databases in crime research is that minorities such as African-Americans are overrepresented in prisons, and findings could be used in a discriminatory manner. This was the reasoning behing the yanking of NIH funding from a conference on “genetic factors in crime” in 1992 at the University of Maryland, with charges of it being a “modern-day version of eugenics” (which is actually timeless).

But times, and technologies, have changed. The 1992 objection to even investigating genetic factors in crime predates the DNA Identification Act of 1994 that led to forensic DNA testing by 1998 – now done in all 50 states. And consider the most notorious recent killing sprees. The perpetrator of the worst attack, at Virginia Tech in 2007, was an Asian, Seung-Hui Cho. The Columbine killers Eric Harris and Dylan Klebold were white, as is James Holmes of the midnight movie massacre in Aurora, CO in 2012. And the blurry, terrifying lone image of Adam Lanza is stark white.

I’m playing Devil’s advocate here. I agree with other geneticists that looking for clues to the Newtown tragedy in DNA could do more harm than good. I also agree that environmental influences on behavior and personality are as important if not more so than inherited factors. But at the same time, I can’t help thinking of those forensic DNA databases and the clues to violent behavior that they may hold, anonymously searchable by crime. And we now have the technology to derive much more information than we did when the technology was limited to selected repeats — we can sequence genomes. That’s a lot of information.

Uses of forensic DNA technology are already eclectic enough to embrace investigation of a criminal tendency profile.

STR typing has been used to identify disaster victims, to reunite Holocaust victims with their families, and to identify kidnapped children. And DNA profiling of footballs from Super Bowl games protects a vulnerable public against sports memorabilia fraud.

So despite lingering apprehension from the history of eugenics in the US in the early twentieth century, the threat of stigmatization, and growing acceptance of genetic determinism as genetic testing and genomes/exome sequencing become more widespread, I’m going to make that modest proposal.

I think that the DNA forensic databases may be important sources of information on the role, if any, of genetics in predisposition to violent behavior.

We have the data. Why not take a look? It’ll keep bioethicists busy for years to come – and might prevent a crime.

  1. Structural or chemical abnormalities in the brain can impair judgment, impair impulse control, cause increased mood instability and aggressiveness, cause obsessive tendencies, impaired empathy, or an impaired reward system (rewarding for socially inappropriate behavior).

    Sometimes these traits can be hereditary, and sometimes they result from injuries or other environmental exposures.

    DNA analysis can help us to find genetic causes. I believe it is more helpful than not to identify genetic causes for these types of problems, because it may help us to improve diagnostics and treatment, and may also help us to reach the understanding that attackers are often victims themselves, suffering from mental illness, and that what we view as a “choice”, sometimes is not so much of a choice, but an episode of severe mental illness. We are, all of us, biochemical robots, with hardware and software, and these events occur when the hardware or the software is malfunctioning. How much choice is really involved? Some, perhaps, but less than is assumed, and perhaps we need to start taking that into account when deciding whether to throw people in prison after the fact, or help them before the fact. At least we can learn more about figuring out who needs help, and what type of help they may need. And, maybe we can even learn how to help them. In many ways, treatment for these types of problems in is it’s infancy, and ineffective in many cases. We need to learn all that we can, so that we can get more effective at helping people who seek help.

  2. Thank you dfwmom. I was hesitant to write this blog, but several people asked me specifically to address the news about Lanza’s DNA. My first kneejerk reaction was to say what everyone else did — that it’s a terrible idea. But then something was nagging at me, and I finally realized it was all that forensic data, and whether or not we should use it. I’d like to see it used to learn more about how to help people, to find out how and why people commit such incomprehensible acts, as you suggest, but it could be a Pandora’s Box situation. It highlights a challenge of genetics in general — what to do with information. But what a tragic example.

  3. I’m sorry to inform you but based on recent statistical analysis of genetic profiles of 100,000 other individual convicted of your crime, combined with your genetic profile, the likelihood a repeat offense if x:1. Therefore we can not grant you parole.

  4. I like your essay … and its “modest proposal.” My major concern is that geneticists and other scholars/investigators may discount, undervalue and overlook important theoretical and technological tools required to assess the critical importance of situational and environmental issues. Far too often, geneticists and others cite correlations and/or associations without understanding that causality is the bottom line. If, for example, standard or novel epigenetic mechanisms are at play (e.g., genetic translation and/or transcription errors [GERRs]), then there could be a tendency to cite genetic issues even though the epigenetic issues may be determinative. Another concern is that there could be a tendency to undervalue diversity issues in healthy individuals (cf. Xue Y et al. 2012. “Deleterious- and Disease-Allele Prevalence in Healthy Individuals: …,” American Journal of Human Genetics 91, 1022-1032).

    For me, a bottom line is to focus on expanding upon our technological tools, both in theory and in practice. In theory, some epigenetic GERRs can be caused by and/or associated with stress-activated viruses (i.e., “autovirulence” associated the Epstein-Barr virus and selected adenoviruses; Smith RW. 1984. AIDS and ‘Slow Viruses’, Annals of the New York Academy of Sciences 437:576-607). In practice, a preliophic moleculator (i.e., PRotonic-ELectronic-IOnic-PHotonIC MOLECUlar calcuLATOR) platform and process specifically were invented to explore autovirulence and other molecular pathological issues (see Smith RW and Shadel RR. 2010. Preliophic Moleculator Using Electric Fields and Gradients for Manipulating Molecules. US Patent #7,826,974 B2 (November 2, 2010).

  5. Excellent point James. I don’t think we will ever be able to include gene expression in any equation, so I doubt that my scenario will become reality. But it would be interesting to perhaps use info from forensic databases in drug development. Restrict use to this application.

  6. Yes it can keep bioethicists busy for years, years they could use to identify other disease genes that would REALLY save lives instead of just stick to speculative eugenics.

  7. One problem with this approach is that STRs run very strongly in families, so much so that one can submit one’s Y chromosome STR profile to web sites such as and get predicted surnames. So any STR marker that correlates with criminality could be explained by the startling conclusion that criminality runs in families, something that would be no surprise to anyone who had watched The Soprano’s. On a more serious note, this means that it may be impossible to control properly for this correlation in a deidentified database, and it will probably not be sufficient to simply use an ethnically matched cohort as a control. And all of this conveniently leaves aside confounders such as environment and culture.

    Even if somehow this could be overcome, one runs into the problem that correlation is not causation, and even if the results are replicated in an independent cohort one is still left with a marker that may function in those two cohorts but fail in others. Without identifying the causative genetic factor it is a virtual certainty that injustice will be done. And how exactly would one identify the causative factor? To my knowledge there is no animal or in vitro model of criminality, and human experimentation is obviously out of the question. Finally, even if somehow one could find the causative genetic factor when would you use it to make decisions affecting human lives?

    All in all, it seems that there are many more important issues to tackle, and just because a set of data is available does not mean it is productive to use it in the most obvious way.

  8. Thanks Darin. I agree with you — I just wanted to bring up a point that I did not see mentioned in the coverage of the Lanza DNA. I wonder what, exactly, the researchers are going to do and look for. The announcement held very little information. Thanks for posting.

  9. Forgive my off-topic ‘pimping’ but: for what little it is worth, my own ‘modest proposal’ re genome sequencing was to make it a condition of filing a complaint of potential ‘vaccine damage’ to have the patient’s DNA sequenced. My idea draws from that there are now a small number of cases rare genetic disorders tackled this way and that there is a reasonable prospect that a number of these vaccine complaints are in fact rare genetic defects. (So that the vaccine complaints are in effect a potential cohort of genetic defects that might be usefully tapped.)

    I apologise for ‘hijacking’ this thread to ‘pimp’ this, but I got no response to my suggestion at the time and perhaps some of your readers might have thoughts on it :-

  10. Grant, thank you for sharing, this is a great post. I’d think a physician would consider rare disease diagnosis when seeing a child with a spectrum of symptoms — but with hysterical parents convinced the symptoms are due to a vaccine, and perhaps a physician not that savvy about rare genetic diseases, I can see where erroneously blaming the vaccine could indeed happen. But once we have newborn genome sequencing — already possible — it will be routine to search for genetic answers. Thanks for posting!

  11. ^^This. The event that inspired the “sequence to find murderer genes” initiative is a pretty obvious case where sociology is likely far more informative*. Let’s think of exactly how these genetic factors, if there are any with large enough effect sizes to matter, would actually influence behavior enough to cause a person to kill a large number of people in a spree. Would it be a simple dysregulation of emotional responses? Not quite, since such acts are quite premeditated rather than a result of ‘lashing out’. Is it a result of autism spectrum? No, no, and no–just ask any autistic person, since there are plenty on the internet and some published (much better sources than armchair diagnoses or neurotypical professionals who have trouble communitcating with the patient in the style the patient is more comfortable with). Is it schizophrenia? Not a consistent causal relationship, see what I said about emotional dysregulation. Tumor in the ventromedial prefrontal (not a genetic factor, of course)? Perhaps that could be a facilitator, but again, loss of inhibitory pathways does not explain what puts these killing acts on the map of “things to do”. Sociopathy? This is a complex phenomenon, and seems to involve a learning process.

    But that is exactly the point–some genetic factors may enhance susceptibility to environmental factors, but you will likely find the social factors more predictive. What if those environmental factors that strongly influence the development of certain pathologies are in turn influenced by a genetic factor, but in a different way? For example, mental illness such as bipolar disorder or schizophrenia are not only worsened by poverty, they can also lead to it when patients are stigmatized and lose their jobs or cannot get them, even if the patients could afford treatment (if they cannot, they more than neurotypicals will be particularly unable to get out of poverty, since they will also have to deal with mental illness along the way). Genetic factors that affect mental illness can also cause a person to be or feel alienated by their peers–but this is as much influenced by peers’ attitudes, which are shaped by the cultural background. When violence, toxic ideas about masculinity, and peer social group dynamics at an early age are mixed in, I just don’t see how much genetics will tell us.

    In fact, depending on how it is sold publicly, this kind of study can easily distract from more predictive lines of questioning and instead foster some very damaging ideas. Racism is one blatant example, as it already severely corrupts the entire criminal justice system. Genetic discrimination, and especially discrimination against mental illness, *will* happen–I have yet to see an example of potential discrimination (against people historically disenfranchised) that did not become pervasive.

    This is coming from a molecular biologist–one who has learned over and over again that it is easy for an expert in one field to underestimate the effects of factors they are not familiar with. I see this when some theoretical physicists claim they can clear up certain issues in biology (say, the entire multifactored condition of cancer, or debates about evolutionary population genetics) with some appeals to thermodynamics, ‘reemergence of ancestral states’, and power laws. I also see this when some biologists studying genetics/developmental/evolutionary biology try to poo-poo all of sociology, as if culture and behavioral history of populations isn’t part of studying a species.

  12. Thanks, Pedram, i agree with you 100%; I’m a geneticist. You beautifully explain the back-and-forth between inherited and environmental factors — they can certainly never be teased apart in any sort of meaningful way. And I agree that discrimination is inevitable, even though the most horrible mass killers are white guys. I wrote this particular blog, against my better judgment, because people asked me my opinion, and I noticed the quoted bigshots dodging the existence of the forensic databases. Can of worms!

  13. Don’t worry, I thought your treatment was one of the more even-handed ones I’ve read. I’ve seen far too many “crazies be crazy” diatribes over the past few weeks, as well as equally unexamined arguments of the “biology has no role in human nature!” strain.

    It is clear you know these nuances well (I’m a regular reader of the blog and I love your insights about genetics; it really puts things in perspective and raises knowledge about the newest breakthroughs–I learn a lot). I was just mentioning some general caveats.

    Actually, rereading my post, I’m afraid I came off as condescending…I certainly did not do a good job making clear that my ire was aimed at the “crime=genetics” crowd. And I didn’t intend it to seem like I’m explaining genetics to a geneticist. Well, since intent isn’t magic, I apologize for that.

  14. One of the issues facing us today is the belief that our actions are often caused by some genetic abnormality. How many labs currently conduct research into obesity genes. However, epigenetics has shown that quite often, the environment or our own behavior is as important. There are two issues concerning DNA data linked to violence and criminality. First, no pattern may be established showing that criminality is more the product of upbringing or environment. Second, that certain genetic patterns may actually be established showing that certain genes could control our behavior. Of course, if the second issue is true then we could be facing a similar future as given in Minority Report. Would there be some form of registration for people with genetic disposition to criminality?

Leave a Reply

Your email address will not be published. Required fields are marked *

Add your ORCID here. (e.g. 0000-0002-7299-680X)

Back to top