Experts Concerned by Use of “HIV Fingerprinting” to Convict

February 15, 2011

In a recent article in the Lancet Infectious Diseases, experts warned that a forensics technique called HIV phylogenetic analysis, sometimes called HIV fingerprinting, cannot definitively establish whether a specific individual transmitted HIV to another person.  The authors also stressed the need for scientists to recognize the limitations of the technique as a basis for proving HIV criminal transmission.

by Kieryn Graham & Courtney McQueen 
The AIDS Beacon

“Phylogenetic analysis is more powerful in its ability to exclude certain scenarios,” said Professor Anne-Meike Vandamme of the Rega Institute for Medical Research at Katholieke Universiteit Leuven and coauthor of the Lancet article, in correspondence with The AIDS Beacon. “Phylogenetics can prove that people cannot have infected each other, but it can never prove that people infected each other.”

The experts listed several guidelines for scientific experts to follow in order to prevent the misuse of HIV fingerprinting evidence in HIV criminal transmission trials. The recommendations include using a sufficient number of comparison samples from other people in the area who have HIV; not informing experts as to which HIV samples are from the accused and which are from the accuser while the tests are being run; and taking special care to correctly word their findings in light of the limitations of the technique (see article). 

HIV Criminal Transmission And Phylogenetic Analysis 

Criminal transmission of HIV is defined as the intentional or reckless infection of another person with HIV. In most states, it is illegal for a person with HIV to have unprotected sex without disclosing his or her HIV status.

HIV phylogenetic analysis is a technique that examines small differences in HIV virus genes in order to determine how closely related two samples of HIV are. In cases of criminal transmission of HIV, it has been used to compare samples taken from the alleged victim and from the accused individual in an attempt to determine whether the accused was the source of the transmitted HIV.

Phylogenetic analysis has been used as part of the overall evidence in several HIV criminal transmission trials in the United States, including a 2004 case in Washington and a 2009 case in Texas. In both cases the accused individuals were convicted.

Last year a scientific article published in the Proceedings of the National Academy of Sciences (PNAS) described how HIV phylogenetic analysis was used to identify the source of HIV in these two criminal cases.

Limitations of HIV Phylogenetic Analysis

The new Lancet article was written partly as a response to the PNAS study. Several experts in phylogenetic analysis cautioned that the technique has the potential to be misused and that by itself, it cannot prove transmission of HIV. 

Unlike human DNA samples or fingerprints, HIV is not unique to an individual. If scientists find a person’s DNA or fingerprints at a crime scene, for example, they can be fairly certain the person was at the scene. 

HIV, on the other hand, is not unique; several people can have very similar HIV, particularly if they were infected by the same person. In addition, even within one person, HIV mutates so rapidly that HIV samples taken at different points in time will look different. This means that no two HIV samples (even within one person) will ever be identical, which makes it difficult to determine the source of a transmitted virus.

Other factors may complicate the matter further, such as the possibility of unknown third parties who may be involved in the transmission of HIV.

“A lot of infected people don’t know that they are infected; the source of a particular infection can be someone who is not diagnosed or someone who is diagnosed but not among the people sampled, which is a big problem,” said Prof. Vandamme.

“There can be a male who gives the virus to another male, who gives it to another,” she explained as an example. “There is one person in between. If you don’t have the virus from the male in between [for comparison], it will lead to the same conclusion as if that first male had immediately infected the third male.”

As a result, the authors argue there can never be complete confidence that the accused individual infected the other person based on phylogenetic analysis alone.

Another problem, says Prof. Vandamme, is that scientists may be giving a false impression of what phylogenetic analysis is capable of.  Prof. Vandamme said she has been in several court cases where the scientific experts from the prosecution failed to carefully communicate what could and could not be concluded from the technique.

“The issue here is the concept of ‘fingerprinting’ as identifying someone as a person who matches the person the court is looking for. But ‘fingerprinting’ really gives the feeling of identifying a perfect match, which you can never do with HIV,” said Prof. Vandamme.

Instead, expert witnesses need to make sure they also include a discussion of the technique’s weaknesses. “If you don’t choose the right wording to express what you can and cannot say you risk giving false confidence. You have to say what you can say, but also stress what you cannot say and that’s not always done properly,” she added.

Debate Within the Phylogenetic Analysis Community

Professor David Hillis, coauthor of the PNAS article published late last year, agreed with the guidelines proposed by the authors of the Lancet article and stressed that he followed similar rules in his own study.  He also stated that in the Washington and Texas cases, there was additional evidence and phylogenetic analysis was not used by itself for conviction.

He also agreed that, “Analyses should not be termed ‘HIV fingerprinting.’  Rather, the correct term is ‘HIV phylogenetic analysis.’”

However, Prof. Hillis does believe that in certain cases, phylogenetic analysis can be used to determine whether one person transmitted the virus to another. “A more detailed analysis of multiple strains from the victim and the suspect, as we conducted in our recently published paper, can indeed support or reject a hypothesis about the direction of transmission between suspected transmission pairs,” said Prof. Hillis.

Prof. Hillis’ analysis relied on the fact that HIV transmission is commonly caused by only a single virus particle, which then mutates and evolves into several different strains. By looking at these multiple strains and how they are related to the HIV in the person suspected of transmitting the virus, Prof. Hillis argues that phylogenetic analysis can infer the directionality of the transmission.

Nonetheless, Prof. Vandamme and many other phylogenetic experts remain unconvinced. “No phylogenetic inference, no matter how confidently done, allows the exclusion of alternative scenarios with an intermediate transmission from one or more unsampled individuals,” she argued. “This method, like any other phylogenetic method in HIV forensics, does not exclude an indirect, rather than direct, link.”

“As forensic experts, we see phylogenetic analysis increasingly being relied on in convictions by criminal courts and with this [PNAS] paper, we fear a renewed false confidence about the reliability of phylogenetic analysis to correctly reconstruct an HIV transmission history, which may lead to risks of miscarriages of justice,” added Prof. Vandamme.

In 2006, criminal charges were dismissed after a judge ruled that phylogenic analysis could not prove HIV-transmission.  A year later, UK experts further explained the process’ limitations.

Tags: , , , , , , , ,

You must be logged in to comment

Log in