News Post

By Samantha C. Spencer

DNA (deoxyribonucleic acid) analysis has long been labeled the gold standard of forensic testing by the court. While I acknowledge the scientific process of DNA testing as the pinnacle of scientific rigor when employed correctly by qualified scientists, certain aspects of the process can be shown to fall short of the objectivity necessary to be considered a scientific standard. To be considered a scientific standard, a method or technique must survive scrutiny and peer reviews. DNA analysis fits into the section of a crime laboratory called forensic biology. This unit is typically separated into two parts: serology and DNA analysis.

Serology is the study of body fluids, and most forensic labs focus on identifying blood, semen, and saliva through screening, color change, and immunological tests. Body fluids are known to contain higher quantities of DNA making the downstream steps of DNA analysis much easier. Once a body fluid has been identified, the item is subjected to DNA analysis which is composed of five steps: extraction, quantitation, amplification, separation, and interpretation. The first four parts of this process are where the science happens. Scientists use chemicals and instruments to produce DNA profiles from evidentiary samples and potentially known samples. These techniques are used throughout the scientific community in research and diagnostic labs.

The final step, interpretation, is unique to the field of forensics, and where the process loses the gold standard label. Developing a mixture profile, or a DNA profile with more than one individual, is extremely common for most evidence submitted to forensic labs. Mixtures are the crux of sexual assault matters, so much so that an extraction technique was developed to attempt to separate the male and female portion of a sample, sometimes referred to as the sperm fraction and the non-sperm fraction. Due to the increased sensitivity of chemistries and instruments and increased requests to sample for trace DNA (not touch DNA – a topic for a whole other conversation), labs see mixtures of two to five individuals in routine casework. Trace DNA refers to the amount of DNA present and does not imply an action that causes DNA to be present on that item of evidence.

For scientists, sensitivity is a double-edged sword. On one side, we have more data, which means we can potentially use this data to identify more perpetrators or provide more information to investigators and the court. More data also equates to better statistical calculations that can provide more weight to our conclusions. This data helps analysts assist in revealing more of the truth using science. On the other side, it leads to many questions about how well we can confidently and accurately identify an individual.

Mixture interpretation – especially of four or more individuals – has always been a subjective aspect of the forensic biology process and fails to survive adequate scrutiny by peers. As the chemistries and instrumentation sensitivities continue to become more sensitive, the subjectivity of interpreting the results increases. This subjectivity is well-documented. One such study was completed by the National Institute of Standards and Technology (NIST) which issued a report on the findings in 2018. In this study, they compared two interlaboratory studies named MIX05 and MIX13.¹ These studies reflect the years in which they were performed, 2005 and 2013, respectively, and compare the protocols and procedures of labs across the United States on how they interpreted mixture profiles. The labs were given multiple DNA mixture profiles of two, three, or more persons (meaning it could be 3 or 4 or 5, or more) and known standards to compare with these profiles. Then each lab provided conclusions based on their standard operating procedures.

If you are still determining whether subjectivity is a part of DNA analysis, here is your answer. For a single lab included in this study, when the responses were reviewed for one scenario in MIX13, 50 percent of the analysis effectively said, “I don’t know” [inconclusive], 30 percent of the analysts said, “He’s not there” [excluded], and 20 percent of the analysts said, “He’s not only in the mixture, but I can exclude greater than 99.9 percent of the population” [match]. This implies that presenting favorable or unfavorable DNA results to a person of interest may depend on which analyst in the lab processes the evidence. Should this be considered a gold standard? This study expresses that this study does not provide a full view of the day-to-day activities within a lab and the results may not have undergone normal casework requirements (such as undergoing a technical review, or multiple analysts working together).

How can we move toward solving this issue if there is no way to establish a set protocol for each and every forensic lab using different chemistries and instrumentation? The best answer we have right now is probabilistic genotyping (PG) software. The MIX05 and MIX13 study states “We are encouraged by the developments of probabilistic software systems…”.¹ PG software uses biological modeling, statistical theory, computer algorithms, and probability distributions to infer genotypes (DNA profiles) and likelihood ratios. The software uses the files generated by the separation step of DNA analysis, or all the data that is used to produce a picture of a DNA profile (called an electropherogram), and deconvolutes the profile. This deconvolution, or process of determining the possible contributors in the profile, occurs using a statistical method called Markov chain Monte Carlo (MCMC) that relies on random sampling to find the genotypes that best explain the composition of the mixture DNA profile. This software removes some of the subjectivity of interpretation by employing validated biological modeling that can include specific data from the lab. However, if this software processes a sample twice, it will not provide the exact same answer but will be clustered around that best-fit answer.

Currently, two well-known companies have produced software that forensic labs are taking advantage of in casework. They both work on the same mathematical principles described above and can deconvolute mixtures of more than five individuals. Both have been presented at trials in multiple states in the US and accepted in admissibility hearings. PG software allows analysts to harness all the data produced in the scientific process and apply well-studied and mathematical sound calculations to provide a more objective, unbiased, and accurate view of mixture profiles. Although this software is a move in the right direction, it is a tool that must be used aptly and after proper training and validation have been completed.

As an attorney, why should you care about this study, PG, and the subjectivity in forensic biology? It comes down to two essential things: voir dire and cross-examination. The findings from this study specifically highlight the need for improved training and striving for consistency.

When preparing for a trial, you must obtain as much information about the analyst as possible, including their training record, proficiency test results, and personnel files, along with the entire case file. If they have issues with their casework, it will be present in these documents. If their training record states nothing about mixture interpretation in the past two to three years, you need to press them on what they are doing to prevent this subjectivity. Ask them if they have participated in any training in PG and if the lab is moving forward to validate and bring on this software for casework. Don’t be afraid to get them to dig deep into their conclusions during cross-examination. Simply reading the report to the jury provides just the tip of the iceberg of the data they collected. Remember, that report is the result of five steps. All forensic scientists should be able to answer questions about their entire process. Furthermore, they should be up to the task of relaying this information in a manner that can be understood by you, the court, and the jury. Even if the lab does not employ PG, they should have a basic knowledge of what this software is capable of and what information it could provide. That is their duty when they climb onto the stand and are deemed expert witnesses. As the attorney, you must ensure they are held to this standard and reveal this subjectivity so that a jury is fully aware of the standards in forensic biology. Although subjectivity will never be fully removed from the forensics field, all criminal law system participants must strive to ensure we decrease it as much as humanly possible.

Endnote

1. Butler, J. M., Kline, M. C., & Coble, M. D. (2018). NIST interlaboratory studies involving DNA mixtures (MIX05 and MIX13): Variation observed and lessons learned. Forensic Science International. Genetics, 37, 81–94. https://doi.org/10.1016/j.fsigen.2018.07.024.

-30-