How do we deal with big genomic data responsibly?

The latest event organised by the Progress Educational Trust (PET), 'With Great Genomic Data Comes Great Responsibility', came at a perfect time given the flood of emails on data protection (GDPR) that most people in the EU have experienced in recent weeks. But PET's event at the Institute of Child Health in London focused on the all-important issue of big genomic data, privacy and how we are dealing with this new responsibility.

Sarah Norcross, director of PET (the charity which publishes BioNews), started the event by remarking that we now live in an era in which an unprecedented amount of data is collected from our daily lives: by companies, while we are shopping or buying train tickets, or if we purchase one of the many genetic testing kits available for anyone who wants to peek into their own genome. The trouble is that once this kind of data is generated, it needs to be dealt with appropriately.

The evening's chair, Timandra Harkness - a science writer, broadcaster and stand-up comedian - introduced the panel of four speakers.

First up was Professor Ann Dalton, director of the Sheffield Diagnostic Genetics Service at Sheffield Children's Hospital. She gave a case study of a female lung cancer patient as an example of how a patient might benefit from genomic testing of their tumour, and how the results might inform treatment choices. However, as Professor Dalton pointed out, 'depending on how much [testing] we do, we generate information that we want and understand, but also information that we don't want at the time or get information we don't really know what to do with at the moment'.

Currently, patients are asked to give consent to have their genomic data collected and many of them are not fully aware of what kind of rights they are signing away, she said. Genomic data is unique in the healthcare context, Professor Dalton argued, as it has an incredibly long shelf life. Test results from an embryo will not only be relevant to the resulting child but to the same person decades later, when they are 80 years old.

Often patients might not want to know additional information that might be uncovered during testing - for example, a diagnosis such as Huntington's disease, that may only affect them in 40 years' time and which cannot currently be cured.

Patients might also not realise that they or their family members can be identified using DNA, and therefore genomic testing does not only affect the individual but their relatives too. 'We don't know what we'll know in the future,' Professor Dalton concluded. She emphasised that it is important to open a dialogue with the public and patients in order to help them understand the significance of their consent signature.

Professor Anthony Brookes at the University of Leicester focused on the usefulness of genomic data in the health context. He emphasised that it is very rare to have a clear gene candidate that is involved in a disease. Usually, the signal within the genomic data is not terribly strong as many different genes or gene networks are involved in disease development, and big data is needed to amplify those signals. At the same time as the signal is amplified, unfortunately, noise can be amplified as well, leading to researchers possibly chasing false leads.

In addition to this, as Professor Brookes pointed out, not all signals apply to the general population and genomic data is not necessarily relevant to a single patient. What it might be useful for, however, is the exploration of human evolution, as well as ancestry and even the identification of criminals.

With some people now voluntarily uploading their genomic data to the internet, it is important to remember that this data might be hacked. It can also be used to identify a person, giving away information about skin and eye colour, inherited conditions and even a likely surname. The challenge, therefore, is to help people understand the amount of information they are giving away, as well as focusing on genomic data as a research tool while being aware of its limitations.

The next speaker, Professor Tim Hubbard - head of genome analysis at Genomics England, the government-established organisation which is responsible for delivering the 100,000 Genomes Project - cast a more optimistic light on the issue. He began with a short introduction into the purpose and structure of the 100,000 Genomes Project, which exploits the dropping cost and shortening time necessary to sequence a genome to collect data from consenting NHS patients with cancer and rare diseases. The idea is to improve the health service on an individual as well as a nationwide basis, and create a legacy of infrastructure for research and treatment that will be useful for future generations.

Selected researchers from outside the NHS can gain access to genomic data from the 100,000 Genomes Project for research purposes, but every effort has been taken to ensure that this data is secure and that individual patients cannot be identified. Hubbard's enthusiasm about the project stood in contrast to the cautionary tales of the earlier speakers - he highlighted the involvement of patient representatives, and the unique opportunity the NHS offers to help us better understand our genomes and make use of this understanding.

The last speaker was Dr Natalie Banner, lead at Understanding Patient Data - an initiative from the Wellcome Trust which seeks to promote better public conversations about how patient data is used. Dr Banner pointed out that 74 percent of people living with cancer are not aware of the UK's cancer registry, and that only 17 percent of people are aware of plans to share their data within the NHS.

The lack of patient knowledge about what happens to their data is problematic, as people are generally receptive to the advantages of sharing their genomic data but only if they are sufficiently informed. By contrast, people become very concerned when given incomplete information, Dr Banner explained. But if we improve the quantity and quality of information provided, there is good reason to expect a high degree of support for the sharing of genomic data.

Dr Banner stressed that this must be a dialogue, as giving consent is not as straightforward as it may seem. We do not know what the future holds and what we might be able to do with genomic data in 20, 30 or 40 years. This is a difficult fact to grasp and work with, and we, therefore, we need to establish a trustworthy and reliable system where genomic data is used for well-defined purposes.

Much of the following panel discussion focused the practicalities of informed consent and how best to establish trustworthy systems, especially in the light of the NHS struggling financially and organisationally.

What all the speakers agreed on, in the end, was that ongoing dialogue with the public is essential.

PET is grateful to Genomics England for collaborating on this event

The next free-to-attend PET/Genomics England event will be 'Whose Genome Is It Anyway? Big Data and Your DNA', taking place in Edinburgh on the evening of Wednesday 10 October 2018.

See here for further details, and email sstarr@progress.org.uk to book your free place.

Films of previous PET/Genomics England events can be watched online here.