Genetic variants are innocent unless proven guilty

variants are innocent unless proven guiltyLast week I described recent data showing that rare genetic variation is a core part of our normal genetic make-up. This includes variants that severely impair gene function. Our understanding of the impact of genetic variation on human health has been fundamentally changed by this new knowledge. The practice of genetic medicine will also be transformed by it.

However, the necessary changes in clinical practice are not happening as fast as they need to. This is partly because how we use genetic testing in healthcare has also fundamentally changed over the last few years.


We used to be very selective about genetic testing

Before DNA sequencing became fast and affordable, genetic testing was very restricted, both in the number of people tested and the number of genes tested. We had to be very, very selective about what we tested and in whom. We typically reserved genetic testing for people with a high chance of having a known genetic condition. I discussed this in more detail in a previous blog.

This rationale for testing heavily influenced how we handled the test results. We considered every identified variant a suspect in causing the person’s condition. And so we exhaustively investigated every variant to evaluate whether it was culpable. Every variant remained a suspect unless it had an iron-clad alibi proving it was innocent, or until the guilty variant was uncovered.

This is similar to interrogating every person at the scene of a robbery. The police initially consider everyone a suspect. After investigations, those with alibis are released. Everyone else remains a suspect until the robber is identified.


We now use genetic testing liberally

Genetic testing is now fast, cheap, and increasingly obtainable. The number of genes and the number of people being tested today is mind-bogglingly higher than a few years ago. We no longer reserve testing for people with a high chance of having a genetic disease, nor for one or two genes. We use genetic testing increasingly liberally, analysing hundreds or thousands of genes to help in diagnosis, management and prevention of disease.

This fundamental change in the scope and rationale for genetic testing should also influence how we handle the results. The traditional variant interpretation paradigm is no longer a good fit. It is neither practical nor suitable.


Genetic variants are innocent unless proven guilty

Traditional variant interpretation processes are not suitable or practical for genetic testing being done today.

The traditional variant interpretation process required us to consider every variant individually to decide if it had an alibi that proved its innocence (a benign variant), if it should remain a suspect (an uncertain variant), or if it caused the disease (a pathogenic variant). This was a reasonable process given the small number of variants to be interpreted and the high chance that one of them had caused the disease.

Now we are testing thousands of genes in thousands of people, many with only a small chance of having a genetic disease. Each test uncovers many thousands of variants, 99-100% of which will be innocent. So we need to switch perspective. Instead of considering variants guilty until proven innocent, we should consider them innocent unless proven guilty.


We need new variant interpretation systems

Extending the robbery analogy may help to explain why this makes sense. The difference between traditional and modern genetic testing is similar to the difference between robbing a bank and cyber theft. A bank robbery usually leads to a small number of local suspects, one of whom is almost certainly the culprit. But in cyber theft the perpetrator could be anyone in the world, anywhere in world. In theory, anyone with access to a computer could be considered a suspect. But if the police told us they were going door-to-door checking computer owners we would be surprised. It would seem a poor strategy, wasting time and resources to rule out people that had little chance of being guilty.

It is not feasible or necessary to pronounce on the likely guilt of every variant identified in a genetic test, any more than it is feasible or necessary to consider the likely guilt of every computer owner when investigating a cyber crime. Instead, we need to develop new variant interpretation strategies and paradigms. Strategies focused on trying to find a disease-causing variant, not on confirming the innocence of thousands of variants that were always likely to be harmless.

Next week I will discuss some of the ways we could adapt variant interpretation processes to achieve this.



Leave a Comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.