By Rachel Horton, William Macken, Robert Pitceathly and Anneke Lucassen.

We’ve each got four to five million ways in which our DNA differs from the ‘standard’ or ‘reference’ genetic code. When our genome is looked at in a test, all these differences are logged. Most won’t affect our health at all, they’re just natural genetic variation, though they could be used to infer stuff ranging from where our recent ancestors probably lived, to what colour our eyes are likely to be. A few of these differences will have a clear impact on health – for example, they might influence our risk of developing cancer, or might mean that if we have children there’s a chance they could have a genetic condition like cystic fibrosis. For lots of these differences, we’ve got very little idea what they mean, or we know that they have slightly different effects in different people.

In getting from four to five million genetic differences per person to a clinical result from a genomic test, a huge amount gets filtered out along the way. This filtering is often seen as a very technical process, a way by which you find ‘the needle in the haystack’ – but that presupposes that we know what results look like and the challenge is simply to find them. In our article we talk through a case where genomic testing done to investigate muscle weakness found something off-target and uncertain: a genetic variation that possibly predisposes to kidney cancer, uterine fibroids and skin lumps. The ‘needle’ we were looking for was an explanation for the patient’s muscle weakness; what we saw instead wasn’t really needle or hay.

The variation in question isn’t causing the muscle weakness that led to the patient having a genomic test, and quite possibly it doesn’t really predispose to kidney cancer at all. It’s completely normal to have stuff that looks at least hypothetically concerning in your genetic code – on average, each person has 54 variations previously reported as disease-causing (i.e. in theory, we’re more confident than we are for the variation discussed here that they might cause trouble) in their genome. In practice, some of these might lead to disease in them or their family, but most won’t. However, in this particular case, a detailed family history asking about kidney cancer, uterine fibroids and skin lumps, and a careful clinical examination focussing on skin, might shed light on the meaning of this particular genetic variation.

So we argue that the variation should be discussed with the patient who had the test – not because it is medical information, but because further work might lead to its becoming so. But we reflect on some of the seemingly technical aspects of genomic testing that led to this variant being considered as a potential result – for example, the choice to examine the gene that the variant is in when exploring the cause of the patient’s muscle weakness, and how the person reviewing the genomic data happened to have a general genetics background rather than an exclusive focus on muscle conditions.

Our article suggests that when appraising genomic information, we should not leap to ask ‘what should we do about this result?’ before we have first questioned why and whether it constitutes a result. Ethical debates around genomics often focus on whether to look for, or how to respond to, genomic ‘results’, but finding results within a person’s millions of genetic variations is not a matter of waving a metal detector around and waiting for needles that make it beep – needle/hay distinctions are often in the eye of the beholder and the choices that need to be made as to why and when genetic variations should be viewed as results deserve more attention.

Paper title: Discussion of off-target and tentative genomic findings may sometimes be necessary to allow evaluation of their clinical significance

Authors: Rachel Horton (1, 2, 3), William Macken (4, 5), Robert Pitceathly (4, 5), Anneke Lucassen (1, 2, 3)

Affiliations:

1. Clinical Ethics, Law and Society (CELS), Wellcome Centre for Human Genetics, University of Oxford, Oxford, UK
2. Clinical Ethics, Law and Society (CELS), Primary Care Population Sciences and Medical Education, University of Southampton Faculty of Medicine, Southampton, UK
3. Centre for Personalised Medicine, University of Oxford, Oxford, UK
4. Department of Neuromuscular Diseases, UCL Queen Square Institute of Neurology, London, UK
5. NHS Highly Specialised Service for Rare Mitochondrial Disorders, Queen Square Centre for Neuromuscular Diseases, The National Hospital for Neurology and Neurosurgery, London, UK

Competing interests: None declared

Social media accounts of post authors: @rach_horton, @w_macken, @RobPitceathly, @annekeluc