In the middle of the night, a woman wakes to find an unrecognizable man, wearing a mask, standing over her. Upon her waking, he forces himself onto her. After assaulting his victim, he flees into the night, leaving the woman with no idea of her attacker’s identity. Authorities have called this man “the Serial Creeper,” and he has terrorized up to 40 women in the Miami, Florida, area. The offender is still at large, and, with no way to obtain a clear sketch of the masked offender from his victims, authorities have turned to a new technology that allows a facial profile to be created through the analysis of DNA.1
In early 2017, Louisiana authorities reported an arrest in a 2009 cold case homicide, based on a facial profile generated through DNA analysis. While DNA was recovered at the scene of the crime, it did not yield a database match. In 2015, the DNA was used to generate a facial profile, and it was a tip from the resulting image that led to the arrest. While the case is still pending at the time of publication, it illustrates a potential law enforcement application of DNA facial mapping. 2
Brief Overview of DNA
DNA (Deoxyribonucleic acid) is a self-replicating molecule present in all living organisms that encodes their genetic information.3 DNA evidence first appeared in the court system in England in 1986, when Alec Jeffreys, a molecular biologist who was studying the use of DNA for forensics, proved the innocence of a 17-year-old boy accused of rape.4 DNA evidence has since become vital in the advancement of criminal investigation.
DNA analysis has become the most significant identification tool for police investigators and forensic scientists since the adoption of fingerprinting at the beginning of the 20th century. DNA allows for individual identification. Other than identical twins, who share the same genetic material, it would be exceptionally rare for two persons to have the same DNA profile.5 DNA analysis can be used to link a perpetrator to a crime, but if the suspect is not in the DNA database of known offenders and if the victim cannot positively identify the suspect for a sample to be collected, law enforcement may need new technology to assist in their investigation. One such emerging technology is DNA facial mapping (DNAFM).
Facial Recognition through DNA Mapping-Overview and Benefits
DNAFM is the ability to predict an individual’s facial features to facilitate recognition, by extracting certain genes from the individual’s DNA. Specifically, scientists can examine selected genes in an individual’s DNA profile to determine which genes have previously been linked to facial characteristics and to consider how normal variations in individuals’ genes could affect the appearance of a face.6
An electronic image of a subject is generated through DNAFM using computer programs. The images produced from this technology are similar to avatars—they are computer-illustrated, but lifelike, images.7 Researchers have identified physical and facial traits suggested by DNA, including eye color, hair color, age, gender, height, genetic heritage, and others, through the use of a system that analyzes 24 genetic variants.8 When facial images are created, percentages of accuracy for each trait are also provided to assist in determining the validity of the results.
The accuracy percentages are measurements of the technology’s confidence in the presence of each physical trait it identifies.9 For example, for one profile, DNAFM technology could be 88.6 percent confident that a male is fair skinned, with an 88.3 percent confidence that the suspect has hazel eyes and a 22 percent confidence that he has freckles. In this example, the facial features with percentages above 80 percent suggest that the technology is very confident that those features will be distinct and present. The percentage involving the freckles is low, which means that the technology reveals that the DNA showed signs that freckles could be present, but that their presence could not fully be determined.10 The intent is less to generate a fully accurate image, but rather it is to illustrate a potential profile. This, along with other investigative information, can be used to generate a suspect profile or to potentially narrow an existing suspect list.11
DNAFM can also generate other benefits. In a case where a mother and her child were murdered and no eyewitnesses or surveillance cameras were present in the vicinity, investigators used DNA taken from the crime scene a to produce a DNAFM-generated image used for a wanted poster.12 DNAFM may also prove to be a valuable investigative tool in age-progression cases, such as for missing children. For instance, if a child goes missing and there has not been a breakthrough in the case for several years, a strand of hair taken from the child’s hairbrush could be used to create an electronic and genetically informed age-progression image. Likewise, DNAFM may be used to supplement traditional methods of facial reconstruction to aid in identification in cases in which skeletal remains are discovered.13
One area in need of further research is the extent to which DNAFM images could be utilized with facial recognition technology. Recently, New York City launched an initiative to track drivers and passengers entering and leaving the city.14 Advanced cameras and sensors are being installed around the city to read license plates and test facial recognition software and equipment.15 With further advancements and research, the use of the avatar-like images produced from DNAFM to match images detected by the camera-based facial recognition systems could potentially be an investigative benefit.
Considerations and Recommendations
Although DNAFM is still in the pilot and testing stages, its use could offer several potential benefits to law enforcement agencies; at the same time, agencies must consider a variety of potential implementation issues. These issues include accuracy, legal concerns, privacy rights, admissibility, and cost.
Scientists must be knowledgeable about factors that affect the results of these images. Although the facial recognition aspect of the technology is driven by software, the second part, facial identification, is human based—and thus open to error.16 Also, while DNAFM can produce images that can serve as a tool for identifying suspects or narrowing a suspect pool, mutated genes, and non-genetic alterations in appearance can affect the accuracy of the production of the image; consequently, this could hamper the identification of the correct suspect, as could DNA from identical twins.
A gene mutation is a permanent alteration in a DNA sequence that makes up a gene and differs from what is found in most people. These mutations can be either hereditary or acquired. If the mutated gene is hereditary, that means an individual was born with the altered sequence. Conversely, an acquired gene mutation occurs at some time in a person’s life, usually through environmental factors, and affects only certain cells.17 For example, some mutations in skin cells can be caused by ultraviolet radiation from sunlight. Alterations may also occur as DNA copies itself during cell division.18 Trained professionals must be able to account for mutated genes when using this technology because acquired mutations will be difficult to detect, thus affecting the results of DNAFM. In addition, an individual’s identity may be different from what DNAFM might predict due to non-genetic alterations in appearance, such as those caused by cosmetic surgery, injury, or scarring.
Finally, as noted earlier, identical twins possess the same DNA sequence. DNAFM can help lead investigators to potential suspects and standard DNA analysis can provide evidence of guilt, but when two individuals share the same genetic sequence, DNA-based results ordinarily might not be dispositive. However, Graham Williams, from the University of Huddersfield in the United Kingdom, conducted experiments on twins’ DNA and discovered that twins can be individually identified genetically by the temperature at which their DNA melts.19 The results of Williams’ experiment concluded that every individual’s DNA melted at a different rate regardless of identical sequences.20 This test can be conducted within a few hours and allows a quicker and more cost-effective process to match a DNA sample to a particular twin.
In the United States, the applicability of DNAFM may be impacted by Fourth Amendment considerations, which protect against unlawful search and seizure. One issue that courts will potentially have to address is whether DNAFM images could be considered reasonable suspicion or probable cause for further law enforcement action, such as investigative detention or issuance of a search warrant pertaining to an individual whose appearance matches the image. The percentage of accuracy and confidence in the appearance of a trait that is produced with each facial image could be instrumental in addressing this question. In addition, DNAFM, by itself, is not sufficient to identify a subject, so the usual standards for making a personal identification will still apply. Once in custody, a DNA sample of the suspect should also be taken and tested to validate the computer-generated image.
Controversy over privacy rights may also arise with the introduction of DNAFM, as it produces facial images from DNA that could potentially be disseminated to law enforcement officials and the public.21 Agencies must clearly outline how DNAFM images can be used and distributed.
Another unresolved question is the extent to which this technology could be admissible in court. Frye v. United States and Daubert v. Merrell Dow Pharmaceuticals set the foundation for judging the admissibility of scientific evidence in court.22 In 1923, Frye v. United States created the “Frye Standard” when the District of Columbia Circuit Court of Appeals ruled that, for scientific evidence to be admissible, it “must be sufficiently established to have gained general acceptance in the particular field in which it belongs.”23 In 1993, Daubert v. Merrell Dow Pharmaceuticals created the “Daubert Standard” and extended the Frye Standard to add that the admissibility decisions must be made on “scientific validity—and thus the evidentiary relevance and reliability—of the principles” of the technology.24 Agencies must consult with their legal advisors to determine whether DNAFM will meet the Daubert Standard.
Agencies will also need to consider the financial cost of this technology. Costs include not only access to software programs and the equipment used to generate DNAFM profiles, but also updates to maintain the currency of technology and training employees in the technology; alternatively, subcontracts with individuals or firms specializing in this technology could be utilized, which would also incur cost. The addition of a new forensic tool can be expensive, so agencies need to determine if DNAFM is a cost-effective option for them.
Based on these considerations, law enforcement agencies utilizing DNAFM will need to create policies to guide its use and application. Agencies must consider their budgets and the potential cost of DNAFM. As DNAFM becomes more widespread, a thorough evaluation process should occur to determine the effectiveness of DNAFM and its benefits to law enforcement.
Conclusion
While tests and pilot applications of this technology continue, DNAFM can prove to be a vital tool in forensic investigations. If law enforcement agencies choose to implement this technology, they should consider both its benefits and limitations. If implemented, DNA facial mapping could help create breakthroughs in criminal cases.
Notes:
1Jim DeFede, “Exclusive: Revolutionary DNA Tech Used To Create ‘Serial Creeper’ Sketch,” CBS Miami, September 10, 2015.
2 KPLC Digital Staff, “Man Arrested in 2009 Killing of Sierra Bouzigard,” KPLC 7, July 24, 2017; KPLC, “Tip from DNA Phenotype ‘Snapshot’ Leads to Arrest in 2009 Murder Cold Case,” Forensic Magazine, July 25, 2017.
3Chris Hardy, “Scientists Use DNA to Predict Facial Features,” Bionews 747, March 24, 2014.
4Lisa Calandro, Dennis J. Reeder, and Karen Cormier, “Evolution of DNA Evidence for Crime Solving – A Judicial and Legislative History,” Forensic Magazine, January 6, 2005.
5Tod Burke, “DNA Analysis: The Challenge for Police,” The Police Chief (October 1989): 92–95. See also Bruce S. Weir, “The Rarity of DNA Profiles,” Annals of Applied Statistics (2007): 358-370.
6Hardy, “Scientists Use DNA to Predict Facial Features.”
7Andrew Pollack, “Building a Face, and a Case, on DNA,” The New York Times, February 23, 2015.
8Ibid.
9Hardy, “Scientist Use DNA to Predict Facial Features.”
10Ibid.
11Kathy Marks, “DNA Phenotyping Shows Results,” Police and Security News (July/August 2017): pages 38–43.
12 Pollack, “Building a Face, and a Case, on DNA.”
13For instance, see Laura French, “Facial Reconstruction Released of Young Girl Found in Suitcase in Texas,” Forensic Magazine, September 15, 2017.
14CS Staff, “NYC Adopts License Plate Readers, Facial Recognition,” Campus Safety, January 13, 2017.
15Ibid.
16 Michelle Taylor, “The Art of Facial Recognition,” Forensic Magazine, March 13, 2017.
17National Institute of Health, U.S. National Library of Medicine, “What Is a Gene Mutation and How Do Mutations Occur?” Genetics Home Reference, March 21, 2017.
18Ibid.
19Jessica Hamzelou, “Police Can Now Tell Identical Twins apart—Just Melt Their DNA,” New Scientist, April 24, 2015.
20Ibid.
21See Christoph Lippert et al., “Identification of Individuals by Trait Prediction Using Whole-Genome Sequencing Data,” Proceedings of the National Academy of Sciences 114, no. 38 (2017):10166–10171.
22Lisa Calandro, Dennis J. Reeder, and Karen Cormier, “Evolution of DNA Evidence for Crime Solving—A Judicial and Legislative History,” Forensic Magazine, January 6, 2005.
23Frye v. United States, 293 F. 1013 (D.C. Cir. 1923).
24Daubert v. Merrell Dow Pharmaceuticals, 509 U.S. 579 (1993).
