Blog post by Shona Kerr, Project Manager of the Quantitative Traits in Health and Disease group at the MRC Human Genetics Unit, University of Edinburgh. She is also Associate Director of the Edinburgh Clinical Research Facility Genetics Core.
Ever since the hype associated with the announcement of the completion of the Human Genome Project in 2003, a “genomics revolution” in healthcare has been promised. Following a decade or so in which evidence of benefits in mainstream medicine has remained elusive, in the past couple of years the speciality of “genomic medicine” has increasingly lived up to the promises made. This is due to a range of factors coming together in a way that is of enormous economic, cultural and societal importance. Chief amongst these are the implementation of efforts (with significant public funding) in many locations worldwide to embed genomics in healthcare systems, exemplified by Geisinger MyCode. They also include the increased availability (and sharing) of research and clinical genomic data internationally, as championed by the Global Alliance for Genomics and Health (GA4GH). Finally, data on millions of members of the general population has been generated by direct-to-consumer genetic testing companies such as 23andMe and Ancestry.com.
Until very recently, clinical genetics data produced in healthcare systems including the UK NHS tended to be small in scale and focussed on a restricted number of genes in which variants were previously known to be capable of causing the clinical condition being investigated. In contrast, most large scale “genomics” data was generated through either publically- or commercially-funded research projects worldwide. At the end of 2018 however, the Genomics England initiative has reached its initial target of 100,000 UK genomes sequenced. Ambitious plans have been agreed for the processes it has established to be integrated increasingly into NHS practice over the next few years. The Genome Aggregation Database (gnomAD) has collected and aggregated genome sequence data from over 140,000 people, with plans to expand to one million genomes over the next three years. This online open access resource has transformed the clinical assessment of which individual variants are too common to cause rare genetic diseases, but currently has no link to clinical measurements from of the people whose genomes have contributed to the resource, restricting its (nonetheless still considerable) value.
The various methods and technologies applied to obtain genomic data in a clinical setting are carefully regulated and performed in accredited laboratories, with all tests implemented for a clear and specific purpose relevant to the health of the person tested. Direct-to-consumer tests of human DNA are also regulated but have somewhat different standards of accreditation and standardisation, with no way for the companies receiving the sample to be able to verify it came from the person submitting it. In a recent entertainment, samples from the pet dog of a journalist were submitted. While most companies reported that the resulting test data was unreadable, one failed to discover that the DNA was not human.
A major challenge is that previously clear lines of demarcation in genomic medicine between “direct to consumer” testing, “research” and “clinical practice” are starting to blur. Large-scale whole genome sequencing is becoming commonplace (it has been predicted that more than 60 million people will have their genome sequenced in a healthcare context by 2025), while more than 12 million people have already had their DNA analysed with direct-to-consumer genetic genealogy tests, and own their data. Very large population research populations or cohorts with genomics data are being created or already exist in several countries. In most instances to date, large genomics research biobank projects explicitly stated in the participant information sheet that genetic test results would not be returned to participants. This includes the UK Biobank, which completed recruitment of 500,000 middle-aged adults across the UK in 2010.
There are good reasons for this strategy. They include the fact that the samples are collected and analysed to research rather than more stringent clinical standards, that sufficient supply of genetic counselling from trained practitioners is not available to help people to understand the significance of their results, and that in normal populations the number of adults with individually useful genetic results was expected to be extremely small. However, communication of “actionable” genetic findings to research participants is now a topic attracting considerable attention. Several influential bodies worldwide are carefully considering which variants in which genes should be fed back to consenting people as additional findings. While this mostly concerns clinical analyses, it can also apply to research projects, if the research governance and ethics approvals of the study allow return of results.
All such “actionable gene” lists only contain disorders that meet a strictly defined list of criteria, namely that all of the conditions can be prevented or treated through surgery, pharmaceuticals or lifestyle changes. Disease gene variants with no clear-cut medical treatments are not disclosed. A range of rates and types of clinically-relevant genetic findings have been reported in different population cohorts, the biggest study of which to date (50,000 people) gave a figure of 3.5% of individuals harbouring deleterious (damaging) variants in a comprehensive list of 76 clinically actionable genes.
The straightforward approach of not communicating results of genetic tests may prove to be increasingly unpopular as the existence of detailed DNA sequence data becomes more widely known to research participants, and its clinical value to a significant percentage of individuals becomes increasingly apparent. In recognition of this, new large research cohorts such as the million people US National Institutes of Health “All of Us Research Program” plan to provide participants with access to results, including genetic/genomic information, according to their preferences, but this also raises a range of challenges. Furthermore, although asking existing cohort participants for new additional consent on return of results is an option, there are difficulties in relying upon consent as a means of ensuring that changing data access practices are rendered ethical, and the right not to know must be respected.
Does it matter whether or not a mechanism is in place for volunteers to know the results of research analysis of their genomes? So long as the explanations were clear at the time of recruitment, the research teams have no obligation to share any of this data with participants, and indeed would often be prevented from doing so by the extent of the approvals granted. But there might be a negative impact on the level of trust in and enthusiasm for medical research if the increasing awareness in the general public of the value of genomic data is not met by mechanisms for the careful return of at least a subset of the most useful results, for which clinical actions to reduce risk of disease can be taken.
With the first tranche of sequences from 50,000 participants in the UK Biobank due to become available for research as soon as spring 2019, there may be a groundswell of excitement from participants wishing to know more about their genomes, and it will be interesting and instructive to see whether or not this likely enthusiasm for knowledge is met. However, it could be argued that if the predictions of the vast numbers of genomes that will be analysed for healthcare or direct-to-consumer purposes in the next few years prove accurate, almost everyone in the UK, either personally or through genomic regions shared with their relatives, will soon have detailed DNA sequence data available, whether or not they have participated in a research cohort.