This article provides evidence for the irrelevance of race-based DNA databases.
The U.S. Federal Bureau of Investigation’s (FBI) core Combined DNA Index System (CODIS) short tandem repeat (STR) panel is required for the calculations of random match probabilities (RMPs) in forensic DNA analysis. Current practice dictates that RMPs should be generated across appropriate reference STR allele frequency databases, including African American, Asian, Caucasian, Hispanic, and Native American, when the suspect’s race is unknown. Should the suspect declare their race, a specific reference database that pertains to that designation is used. This practice is based on the presumption that racial population group is relevant for calculating conservative RMPs that favor the defendant. The core CODIS panel has been expanded to 20 STRs; however, the relationship between RMP and race has not been re-evaluated. The current article reports that genetic structure analyses and Bayesian-based population assignment of expanded CODIS profiles from one race-neutral and five race-specific reference databases revealed that STR data could not distinguish races as distinct biological clusters. For instance, although the average race-specific RMPs for Hispanic or Caucasian profiles were almost equally conservative when calculated from either population's reference database, the Hispanic profiles closely affined with the Native American population. Race-neutral RMPs computed with a correction factor (θ) of 0.03 favored the defendant as much as race-specific RMPs based on a θ of 0.01. Insufficient genetic differentiation observed among the U.S. racial populations, as well as inconsequential differences between race-specific and race-neutral RMPs undermine the value of using “race” in the context of forensic DNA analysis and support the argument that forensic databases should be race-neutral. (publisher abstract modified)