NIST Corner: Complex DNA Mixtures
Written by Dr. Michael D. Coble   

ONE OF THE WORLD’S TOP forensic biologists, Dr. Peter Gill, once remarked in a presentation on DNA mixture interpretation, “If you give ten colleagues a mixture, you will probably get ten different answers.” His comments highlighted the state of mixture interpretation in April 2005 and the lack of consistency and standards for the community. Indeed, DNA mixture interpretation is one of the greatest challenges facing the forensic DNA field.

There are many reasons why DNA mixtures—the presence of two or more individuals in an evidentiary profile—are complex to interpret. One is that for years, forensic biologists have had limited guidance. Prior to the publication of the 2010 SWGDAM (Scientific Working Group for DNA Analysis Methods) Autosomal DNA Interpretation guidelines, the 2000 SWGDAM document provided only seven sentences on mixture interpretation. Secondly, although STR multiplex kits are becoming more sensitive with improved chemistries, ideal DNA profiles often require greater than 250 picograms of input DNA to avoid stochastic effects (that is where not all of the alleles present from the minor contributor are amplified to produce adequate signal in the electropherogram).

Photo: University of Michigan School of Natural Resources & Environment

Multiple studies in the literature have shown that the DNA extraction process is inefficient and only recovers about 20-30 percent of the DNA present in the evidence. This loss of DNA can push what was a high-quantity component of the mixture into a low-quantity component and lead to stochastic effects in the data. Also, in complex mixtures, it may be difficult to know exactly how many contributors are present in the evidence due to allele sharing. Plus, forensic DNA laboratories are analyzing more and more challenging samples such as “touch DNA” associated with burglaries and testing of handguns where multiple individuals have handled the items of evidence. Although this is not an exhaustive list, there are many confounding factors that make low-level, complex mixtures a leading cause of headaches for the forensic biologist.

The Applied Genetics Group at NIST continues to provide resources to the forensic DNA community for mixture interpretation. Presentations and materials from previous workshops with collaborators are available in the DNA mixture interpretation section of STRBase along with references from the literature on the elements of mixture interpretation. In 2013, NIST hosted a mixture interpretation webcast for DNA analysts that is archived at one of the links listed below. Other webcasts on probabilistic genotyping have been presented, and more webcasts are in the works.

In addition to training resources, NIST is conducting research on the current state of DNA mixture interpretation in the United States through an interlaboratory study: MIX13. Since 1997, NIST has conducted five DNA mixture interlaboratory studies. Since the last interlaboratory study (MIX05), several laboratories have refined their protocols and established thresholds according to the 2010 SWGDAM guidelines. With 108 laboratories participating in MIX13, this study will be one of the largest such investigations to date. The goals of MIX13 are to determine the current “lay of the land” of STR mixture interpretation across the community, gauge the consistency in mixture interpretation across the United States, and identify future training and research needs that could help to improve mixture interpretation and reporting. The analysis of this vast dataset is underway with the goal to publish and present the findings to the community.

Another area of investigation at NIST is the usefulness of probabilistic genotyping software programs for complex DNA mixture interpretation. Currently, laboratories interpret mixtures using either the “Random Man Not Excluded” (RMNE) approach where the alleles present in the profile are used to generate a statistic such as the Combined Probability of Inclusion (CPI) or use an “Inferred Genotype” approach where the mixture profile is deconvoluted into individual components and a statistic, such as the modified Random Match Probability (mRMP) or the binary Likelihood Ratio (LR), is used to assess the weight of the evidence to the suspected contributor.

The current strategies suffer when alleles are lost in the profile due to a failure of the DNA to be amplified to generate a signal on the electropherogram (known as allele drop-out). Recently, the International Society for Forensic Genetics published recommendations for the interpretation of low-level mixtures when dropout is possible (Gill et al. 2012). A number of software programs are now available that incorporate a probability of dropout in the LR using either a semi-continuous approach (where the alleles present in the profile are used in the statistical calculation) or a fully-continuous approach (where additional information such as peak height, stutter, mixture ratio, etc… are used to determine the most optimal genotypes in the mixture). The Applied Genetics Group is in a unique position to investigate a number of these software programs that are both freely and commercially available to the community.

NIST scientists have produced standard reference materials and evaluated new technologies for the forensic DNA community for more than 20 years. Ongoing efforts in forensic DNA are part of a more recent NIST-wide focus on forensics, and we continue to share presentations and information about our research on STRBase. Forensic biologists can sign up to receive email alerts from NIST on developments in DNA mixture interpretation and other forensic science topics by going to: and entering an email address in the sign-up box on the left side of the page

About the Author

This e-mail address is being protected from spam bots, you need JavaScript enabled to view it is a forensic biologist with the NIST Applied Genetics Group where he conducts research in forensic DNA testing and clinical genetics.


Mixture Interpretation

Mixture Training for Analysts

Mixture Webcast on Probabilistic Genotyping


Training Materials

< Prev

Lifting Latent Fingerprints from Difficult Surfaces

ALMOST ANYONE can find, process, and lift a latent print that happens to be in a logical and obvious place like a door handle, a beer can, or a butcher knife. But sometimes, a latent print is not just sitting there in a logical and obvious place. Sometimes, you have to use your imagination to find the print and your skills to lift it.