How Do DNA Databases Work?
Renowned forensic scientist and professor Dan Krane, a frequent contributor to the “Hidden Killers” podcast hosted by Tony Brueski, joined the show to shed light on the complexities of DNA in criminal investigations. Known for his ability to break down the intricacies of DNA and forensic science, Krane delved into the realities of DNA databases, their potential errors, and the potential for improvements.
The National Offender Database, which held 7 million DNA profiles in 2009, now houses over 13 million. The exponential growth of the database presents the opportunity to identify error rates in DNA test results. Krane shared an experience from his tenure with the government, where he identified an error rate of one in 300 when he worked on a database of 15,000 DNA profiles.
Krane explained, “When you’re talking about chances of a coincidental match in the quintillions or the bazillions… and yet you make a mistake every 300th time that you try to generate a DNA profile, the bazillion number doesn’t matter anymore.” The implication is that the system could potentially miss opportunities to generate useful investigative leads. For Krane, this represents an “unacceptably high failure rate,” given the gravity of crimes that could go unresolved.
When asked how to correct such errors, Krane suggested the potential for identifying specific locations in the DNA sequence that might be more prone to errors. Excluding these locations when doing database searches could improve accuracy. Moreover, if the system were more transparent, more flaws could be identified and corrected.
However, the current level of secrecy surrounding the database is a concern for Krane. He pointed out the security surrounding the CODIS database, which is tighter than the Pentagon maintains. He stressed the need for more transparency, arguing that it could allow more opportunities to generate investigative leads and solve more cases.
Krane also sheds light on the advanced expert systems, or probabilistic genotyping systems, which interpret complex DNA test results that human analysts find challenging. These systems, utilized worldwide for about five to ten years, undertake billions of calculations, distinguishing between signals and noise in data, which is particularly useful for low-level samples or samples with multiple contributors.
When asked if these expert systems utilize AI, Krane admitted it’s difficult to say. The systems are complex, supported by hundreds of thousands of lines of computer code. As he revealed, the creators of these systems have been less than forthcoming about their workings, often requiring stringent nondisclosure agreements.
Want to listen to ALL of our podcasts AD-FREE? Subscribe through APPLE PODCASTS, and try it for three days free: https://tinyurl.com/ycw626tj
Follow Our Other Cases: https://www.truecrimetodaypod.com
The latest on Catching the Long Island Serial Killer, Chad & Lori Daybell, The Murder of Ana Walshe, Alex Murdaugh, Bryan Kohberger, Lucy Letby, Kouri Richins, Justice for Harmony Montgomery, The Murder of Stephen Smith, The Murder of Madeline Kingsbury, and much more! Listen at https://www.truecrimetodaypod.com
The National Offender Database, which held 7 million DNA profiles in 2009, now houses over 13 million. The exponential growth of the database presents the opportunity to identify error rates in DNA test results. Krane shared an experience from his tenure with the government, where he identified an error rate of one in 300 when he worked on a database of 15,000 DNA profiles.
Krane explained, “When you’re talking about chances of a coincidental match in the quintillions or the bazillions… and yet you make a mistake every 300th time that you try to generate a DNA profile, the bazillion number doesn’t matter anymore.” The implication is that the system could potentially miss opportunities to generate useful investigative leads. For Krane, this represents an “unacceptably high failure rate,” given the gravity of crimes that could go unresolved.
When asked how to correct such errors, Krane suggested the potential for identifying specific locations in the DNA sequence that might be more prone to errors. Excluding these locations when doing database searches could improve accuracy. Moreover, if the system were more transparent, more flaws could be identified and corrected.
However, the current level of secrecy surrounding the database is a concern for Krane. He pointed out the security surrounding the CODIS database, which is tighter than the Pentagon maintains. He stressed the need for more transparency, arguing that it could allow more opportunities to generate investigative leads and solve more cases.
Krane also sheds light on the advanced expert systems, or probabilistic genotyping systems, which interpret complex DNA test results that human analysts find challenging. These systems, utilized worldwide for about five to ten years, undertake billions of calculations, distinguishing between signals and noise in data, which is particularly useful for low-level samples or samples with multiple contributors.
When asked if these expert systems utilize AI, Krane admitted it’s difficult to say. The systems are complex, supported by hundreds of thousands of lines of computer code. As he revealed, the creators of these systems have been less than forthcoming about their workings, often requiring stringent nondisclosure agreements.
Want to listen to ALL of our podcasts AD-FREE? Subscribe through APPLE PODCASTS, and try it for three days free: https://tinyurl.com/ycw626tj
Follow Our Other Cases: https://www.truecrimetodaypod.com
The latest on Catching the Long Island Serial Killer, Chad & Lori Daybell, The Murder of Ana Walshe, Alex Murdaugh, Bryan Kohberger, Lucy Letby, Kouri Richins, Justice for Harmony Montgomery, The Murder of Stephen Smith, The Murder of Madeline Kingsbury, and much more! Listen at https://www.truecrimetodaypod.com
