DNA evidence is one of the primary tools used during forensic analysis. DNA is the set of molecules containing our genetic code and is 99% identical among all humans. The remaining 1% can be different enough to identify individuals and relationships. Identifying an individual can be done through a process called DNA profiling (also known as genetic fingerprinting).
To create a DNA profile, DNA is isolated from a biological sample (e.g., blood, saliva, hair, or skin) and then analyzed for distinct genetic patterns that differ among humans. These patterns are then compared between samples in order to determine the likelihood that they could have come from the same person. When comparing samples, statistical analysis is used to assess if enough similarities occur to indicate the samples likely originated from the same person, or if a familial relationship exists between the two samples.
As differentiated in the PCAST report, forensic DNA evidence falls into two categories:
1. DNA analysis of single-source and simple-mixture samples: Single-source samples are biological samples that contain material exclusively from one individual. Simple-mixture samples, on the other hand, contain biological material from two individuals, one known and one unknown. In the case of simple-mixture samples, the alleles of the known individual can be “subtracted” from the total sample in order to create a DNA profile of the unknown individual. This is a largely objective and routine process with a straightforward statistical analysis. Errors do occur, but the chances of an incorrect identification, according to the report, are close to 1 in 1 billion if the technician is well trained.
2. DNA analysis of complex-mixture samples: Complex-mixture samples are biological samples that contain material from multiple individuals in unknown proportions. A fundamental difference between analysis of complex- and simple-mixtures is that the procedure for complex-mixtures involves a higher degree of interpretation by the technician. According to the report, due to the complexity of DNA and the variations that could be present at any given genetic locus, it can be extremely challenging to separate out distinct individual DNA profiles when considering multiple unknown sources of DNA. The statistical analysis to determine the probability of an individual’s DNA being present in the sample currently involves subjective calls on which alleles to assign to which DNA sample source; changes in statistical analysis tools and subjective calls can lead to vastly different results. The accuracy of identifications and exclusions made by DNA analysis of complex-mixtures is sometimes contested in court and widely variable.
One of the major issues brought up in the report is that prosecutors are presenting DNA results claiming high validity based on the statistical analysis that is applied in any given case. The general consensus of the report is that, in most cases, complex-mixture analysis has minimal foundational validity and statistical analysis methods are highly subjective.
The report concludes that the DNA analysis of complex-mixture samples relies upon subjective review of data and lacks scientific foundation. Although promising, the use of software to resolve mixtures, such as for the 2016 trial of Oral Hillary, has suggested that automated methods are inconsistent and not yet ready for widespread use. The report also references the potential use of probabilistic genotyping algorithms to address mixture resolution.