Forensic Phenotyping: What DNA Can (and Cannot) Tell Us About a Criminal’s Appearance.

DNA in a test tubeAll Points Bulletin: Wanted for Murder: A Redheaded, Blue-Eyed, Left-Handed Smoker Who Likes to Ski, Has an Elevated Risk of Cancer, Is Allergic to Cashews, and Has a Birthmark the Shape of Wisconsin

OK, the thought of issuing such a specific physical description of a suspect seems ridiculous to us now, but can we expect to see such specific descriptions of alleged criminals in the future?

A new field of forensic DNA analysis, forensic phenotyping, is emerging, and it is raising some good questions. How much information can or should forensic analysts glean from someone’s DNA, and how much of that should be made public?

With the recent popularity of crime scene investigation shows on television, I would venture a guess that most of us know a thing or two about the use of DNA in forensic investigations. Granted, much of the information in these TV shows is not exactly accurate. (See The Reality of Crime Scene Investigation. Part I: Common Myths.) However, I think that most viewers get the basic gist of it: Analysis of variable sequences within a person’s DNA can be used to place a person at a crime scene.

Briefly, regions of polymorphic DNA consisting of repetitive DNA sequences are scattered throughout the human genome. In the early 1980s, these regions were investigated as informative markers to map the human genome (1, 2), but one of the founding fathers of DNA typing, Alec Jeffreys, quickly realized that these markers could be used for DNA-based human identification (3). In modern forensic laboratories, the polymorphic DNA of interest is usually short tandem repeats (STRs), and forensic analysts can generate an STR profile from biological evidence left at a crime scene. A match between this DNA profile and a suspect’s profile can identify the “DNA donor” with >99.99% certainty. DNA also can be used to exonerate an accused individual. These types of analyses have been going on for decades and have implicated and exonerated hundreds of thousands of people accused of crimes. Most STRs are located within noncoding regions of DNA, so an STR profile cannot tell us anything about a person’s appearance. Not much controversy here.

However, scientists are now investigating single nucleotide polymorphisms (SNPs) for the sole purpose of generating a description of an alleged criminal. Some SNPs can be used as ancestry informative markers (AIMs) to provide information about a person’s ethnicity. Other SNPs can predict hair, eye and skin color. In fact, commercially available kits already exist to predict red hair color, eye color and ethnicity. One limitation of these tests is that environmental and nutritional conditions can complicate final phenotype determination. Plus, phenotypes can be masked: Someone can have grey hair or shave their head (or be naturally bald).

Side note: A less controversial application of these SNPs is to characterize the ethnicity of ancient remains, including 25 sets of Bronze- and Iron-Age human remains found in south central Siberia (4), and determine that some Neanderthals probably had red hair and fair skin (5).

Currently, the information that scientists can gather from a person’s DNA is limited and often restricted to physical appearance such as hair and eye color and skin pigmentation. The best studied example of genotype determining phenotype involves the melanocortin-1 receptor (MC1R). MC1R is located within melanocyte cell membranes, and binding of α-melanocyte stimulating factor to MC1R increases synthesis of black and brown pigments (melanin and eumelanin). Mutations in the mc1r gene interfere with eumelanin and melanin production, allowing the red and yellow pigments (pheomelanin) to predominate. There are more than 30 known MC1R variants, three of which are associated with red hair, fair skin and freckling (7,8). Population studies have shown that homozygosity or compound heterozygosity for one of these variants confers a >90% chance of having red hair (9). A genetic test to detect these variants is already being used in the United Kingdom as an investigative tool. Scientists also are investigating SNPs to predict eye color and have identified a number of promising candidates (10,11).

In theory, scientists could develop genetic tests that allow future law enforcement officials to detect traits well beyond physical appearance. Could predictive tests be developed for behavioral traits? Traits such as impulsiveness? Aggression? Violence?

Few countries or states have laws governing the use of DNA to determine phenotype (reviewed in 12). In some jurisdictions, only noncoding DNA may be analyzed, and phenotypic information cannot be gathered, with a common exemption for gender assignment. The Netherlands is the only country that explicitly allows forensic phenotyping, within defined guidelines. In several parts of the world, including the UK, one of the leaders in the use of forensic DNA analysis, forensic phenotyping is implicitly allowed.

Koops and Schellekens argued in a recent review (12) that the degree to which forensic phenotyping should be allowed depends on the sensitivity of the information revealed. Analysis of SNPs that influence physical appearance is less likely to create a controversy than analysis of SNPs that influence more sensitive information such as susceptibility to disease, tendency toward aggressive behavior, or homosexuality. In the Netherlands, only externally perceptible traits can be used, and only when that trait can contribute to the criminal investigation. Also, the DNA donor’s right not to know the results must be recognized. In a medical setting, a person has the right not to know about a genetic predisposition to some medical condition. Does that same right apply in criminal investigations? Does an individual’s right not to know trump the public’s interest in apprehending a criminal?

What if the revelation of a trait could cause psychological trauma to the person being tested? A suspect could suffer an identity crisis if an AIM analysis predicts that he has ancestors of an ethnic group against which he is prejudice. Koops and Schellekens (12) describe a theoretical situation where a physical description of a suspect could cause that person to question his paternity: If the person has red hair but there is no history of red hair in his family.

Another concern is the communication of that information to third parties. In cases where a particular phenotype is rare in a community, law enforcement could discreetly contact those people who fit the description and do their investigation without making the profile public. However, in cases where suspects cannot be identified by other means, a physical description of a suspect is often broadcast to the public. It is unlikely that most externally perceivable traits will be a cause for privacy concerns. However, a genetic predisposition toward some undesirable trait, such as pedophilia, aggresiveness or a medical condition, could be used against someone by, for example, a family member, medical insurer or future employer. In communities with certain beliefs, the revelation that a husband and father might be suppressing a homosexual tendency could destroy or alienate his family.

Who makes decisions about what traits are considered “too sensitive” to be made public?

Making certain information such as ancestry available to the public could cause discrimination and stigmatization, particularly if that ethnic group is a minority in that society. Also, existing prejudices within the community could be reinforced, especially if the same ancestry informative markers were found in a string of recent, unrelated crimes. Despite the fact that many global communities are actively trying to overcome and eliminate prejudices, the reality is that prejudice still exists in some form or to some degree in almost all societies. Could a prejudice, whether real or perceived, even extend to law enforcement agencies? One could argue that these risks are no greater for physical descriptions created through DNA analysis than for those generated by an eyewitness account. Does it matter how this phenotypic information is gathered, whether it is from DNA, an eyewitness account or some other means? Would simply omitting information about how a suspect’s description was developed alleviate some of the concerns about DNA-based phenotyping?

I wasn’t sure if I wanted to mention this final point because it seems a little scary, too “Big Brother”, but here goes: Would genetic testing be performed for reasons other than a specific criminal investigation? How would we as a society treat people who are revealed to have a genetic propensity for an undesirable trait? Would such testing lead to pressure on certain individuals who test positive for a trait but do not exhibit that trait to undergo treatment? For example, if scientists discover DNA sequences associated with pedophilia, should people with those DNA sequences be chemically castrated to curb the risk of harm to children, even if those people have displayed no tendencies toward pedophilia? This seems extreme to most of us. Mr. Spock would argue that “The needs of the many outweigh the needs of the few”, but outside of Vulcan society, how acceptable is this? Is there a risk that would-be parents would test their unborn child for behavioral or physical characteristics similar to the way that prenatal screenings are now performed to detect certain genetic diseases? What are the consequences of those tests? Once you know something, you cannot “unknow” it.

When used for its original intent, to develop a physical description of a criminal suspect, forensic phenotyping has some definite benefits in certain circumstances. However, I think we must acknowledge that there are potential risks, especially if the information is gathered and disseminated with few guidelines. Knowledge is power, and we must use it wisely.

What do you think? Would you sleep better knowing that law enforcement has this powerful new tool to help apprehend criminals, or do you think that using DNA to develop a description of a suspect is going too far? Do we allow testing for some traits but not others? Where do we draw the line?

  1. Wyman, A.R. and White, R. (1980) A highly polymorphic locus in human DNA. Proc. Natl. Acad. Sci. USA 77, 6754–8.
  2. Jeffreys, A.J., Wilson, V. and Thein, S.L. (1985) Hypervariable ‘minisatellite’ regions in human DNA. Nature 314, 67–73.
  3. Jeffreys, A.J., Wilson, V. and Thein, S.L. (1985) Individual-specific ‘fingerprints’ of human DNA. Nature 316, 76–9.
  4. Bouakaze, C. et al. (2009) Pigment phenotype and biogeographical ancestry from ancient skeletal remains: Inferences from multiplexed autosomal SNP analysis. Int. J. Legal Med. 123, 315–25.
  5. Lalueza-Foc, C. et al. (2007) A melanocortin 1 receptor allele suggests varying pigmentation among Neanderthals. Science 318, 1453–5.
  6. Valverde, P. et al. (1995) Variants of the melanocyte-stimulating hormone receptor gene are associated with red hair and fair skin in humans. Nat. Genet. 11, 328–30.
  7. Bastiaens, M. et al. (2001) The melanocortin-1 receptor gene is the major freckle gene. Hum. Mol. Genet. 10, 1701–8.
  8. Jobling, M.A. and Gill, P. (2004) Encoded evidence: DNA in forensic analysis. Nat. Rev. Genet. 5, 739–51.
  9. Grimes, E.A. et al. (2001) Sequence polymorphisms in the human melanocortin 1 receptor gene as an indicator of the red hair phenotype. Forensic Sci. Int. 122, 124–9.
  10. Liu, F. et al. (2009) Eye color and the prediction of complex phenotypes from genotypes. Curr. Biol. 19, R192–3.
  11. Valenzuela, R., Henderson, M., Walsh, M., Garrison, N., Kelch, J., Cohen-Barak, O., Erickson, D., John Meaney, F., Bruce Walsh, J., Cheng, K., Ito, S., Wakamatsu, K., Frudakis, T., Thomas, M., & Brilliant, M. (2010). Predicting phenotype from genotype: Normal pigmentation. Journal of Forensic Sciences, 55 (2), 315-322 DOI: 10.1111/j.1556-4029.2009.01317.x
  12. Koops B.J. and Schellekens, M. (2008) Forensic DNA phenotyping: Regulatory issues. Columbia Sci. Technol. Law Rev. 9, 158–202.
The following two tabs change content below.

Terri Sundquist

Terri has worked as a Scientific Communications Specialist at Promega Corporation for more than 13 years, and prior to that, spent more than 5 years solving problems and answering questions as a Promega Technical Services Scientist. She graduated with B.S. degrees in Chemistry and Biology at the University of Wisconsin—River Falls, then earned her M.S. in Molecular Biology from the Mayo Graduate School in Rochester Minnesota.


  1. Personally, I find the fact that hair color is one of the easiest targets of forensic phenotyping interesting. As a redhead, I now have one more reason to stay out of trouble. However, considering that only 1–2% of the world’s population has red hair, I have to ask how useful is a DNA-based test to detect red hair?
    After investigated the topic a bit more, I was surprised to learn that there is quite a bit of variation in hair and eye color in certain parts of the world, even in European countries where the residents are stereotyped as blonde-haired and blue-eyed (see It’s not surprising then that European countries are leading the research in this application.

  2. No I don’t think I will sleep better. Catching a bad guy doesn’t mean that the neighborhood is safe already. Even though I admit that it is better to know that someone who has done and possibly do the same crime again is captured.

  3. Thanks for this, Terri. It’s a slippery slope. One thing I wonder about, is how much easier it is for people of European descent to change their hair or eye color without it looking strange compared to those of African or Asian descent. Your red hair might make you try to stay out of trouble, but it won’t keep you out of the interrogation room if a red-headed suspect pops up.

    1. Good point Karen. I think we’ll be seeing a lot of discussion on this topic in the next few years, especially as scientists identify more biomarkers associated with certain phenotypes in the future. Hopefully, society can agree on guidelines so that genetic information can be used to benefit criminal investigations without violating anyone’s privacy.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.