Google+ Facebook Twitter Twitter

The 100,000 Genomes Project: it's not just about DNA

Gerry Thomas, Professor of Molecular Pathology, casts a critical eye over the 100,000 Genomes Project.*

Genomes DNA sequence

In 2012, the UK government announced the establishment of the 100,000 Genomes Project. It is billed as a project that would revolutionise patient diagnosis and treatment, offering the prospect of personalised treatment for many patients. The project is run by Genomics England, a company that is wholly owned by the Department of Health, but is delivered through 13 newly established Genomic Medicine Centres, comprising a lead NHS trust and a series of local delivery partners – usually smaller NHS trusts geographically close  to the lead organisation.

100,000 Genomes Project breakdown in numbers

The key aims of the project are to promote genetic research in order to bring benefit to NHS patients and to support the development of the UK genomics industry. Sequencing is carried out by a single provider, Illumina, at a purpose-built facility in Cambridgeshire. The programme has two strands – rare disease and cancer. It is designed to be a transformative project and to make interpretation of our DNA sequence sit alongside conventional diagnostic procedures and tests to inform the appropriate clinical pathways for treatment of disease.

For rare disease, only a blood sample is required – usually from the patient and their parents, and most of these families are already aware that a germline genetic component to their disease is a likely cause. For cancer patients, a paired sample of blood and tissue is required. It is this second arm of the project that is more challenging to deliver both practically and ethically.

The aims of the project are laudable and given the recent announcement of investment by big pharma in the UK, the project is already bearing fruit. The old adage of “right treatment to the right patient at the right time saves money” should encourage all of us who fund the NHS through our taxes to engage with the project. However, the delivery of this project is not easy and will involve substantial changes in the skills needed in the histopathology workforce, as well as changes to workflow through the departments. It also raises important ethical issues.


Challenges for pathology departments

Early pilot work carried out in the project suggested that FFPE tissue produced a number of sequencing artifacts and produced noisy profiles for the bioinformaticists to work with. This would predictably lead to the need for more validation of genetic alterations to determine what was real and what was noise, and make translation into the clinical arena more challenging. A decision was therefore taken to use only frozen tissue samples. This is a sea-change in pathology departments that for years have worked to optimise fixation and processing protocols to ensure that downstream diagnostic patterns can be identified using, for example, immunocytochemistry.

The provision of frozen samples means that material must be transferred from the operating theatre to the pathology department rapidly and the pathologist must be on hand to perform cut-up with minimal delay. While this would be practical with onsite pathology departments, the amalgamation of pathology services that has taken place means that histopathology services are often centralised and sometimes at different geographical locations from operating theatres. Transfer of specimens would need to be carried out at 4°C to prevent degradation of protein (for later immunocytochemistry) and DNA/RNA. This raises issues of possible cross-infection and has a cost implication regarding couriers using temperature controlled transport.


Ethical dilemmas

The timing of consent also raises ethical issues. Until the use of WGS is proven to be of diagnostic value, pathologists are quite rightly likely to take frozen samples only when this will not have deleterious effects on the diagnostic process. Where patients are consented prior to their operation, it cannot be known whether the pathologist will be able to take a frozen sample safely, and therefore whether the patient would be eligible for the 100,000 genomes project. It might be considered to be more ethical to only take the detailed consent for the project when the eligibility of the patient is known. The lack of availability of suitable samples (either with respect to biological format or amount) has been shown to be a significant issue in stratified medicine trials such as CRUK’s Stratified Medicine 2 project. Cancer patients are made aware that they themselves are unlikely to benefit from the project; but there is still expectation that they will be entered into the project. Few centres take routine generic consent that would enable retention of tissue for later enrolment in clinical studies that require frozen samples, and few histopathology departments have the capacity to store frozen tissue as part of the diagnostic record.

There are other ethical issues. Better diagnosis through improved imaging means that we are identifying smaller cancers, and these are the least likely to provide material that can be frozen without a potential effect on the diagnostic process. Does this mean that those with small tumours will denied access to the benefits of whole genome sequencing? Will the cost of transferring samples from small hospitals that have no local histopathology facilities mean that access to genomic medicine will become a postcode lottery?

There are also repercussions for the science of the project. The requirement for frozen samples biases the dataset towards larger tumours – can we apply what we learn from this to guide patient treatment for those with smaller tumours? Inevitably, only a small sample of the operative specimen will be taken for sequencing – is this representative of the whole tumour? If not, how many samples are required to counteract the inevitable bias of tumour heterogeneity? How does this impact the cost of delivering a meaningful genetic sequence on which to base future therapy decisions?

Identification of drivers of disease and indicators for treatment will only come from pooling both the genetic sequence data with large and detailed clinical datasets. In the absence of a unified electronic patient record that mandates collection of defined clinical datasets, this is probably even more of a significant challenge than those facing histopathology. The need to link electronic patient records with genetic information may sit uncomfortably with some – but doesn’t Dr Google know more about you from the way you share your life on social media than can ever be gleaned from your genetic code and medical record? Can your political views or sporting allegiances be gleaned from your medical record and DNA?

The delivery of the promise of better healthcare will come at a cost to our autonomy. We must recognise that consent taken at the time of sample donation cannot cover all possible outcomes, as technology changes so quickly that all potential uses of our data and samples cannot be foreseen. Consent, therefore, can never be completely informed at the time it is taken.


Feedback of data

Considerable thought will need to be given to what information needs to be given back and to whom. In taking part in the 100,000 Genomes Project participants agree to receive information back that is directly relevant to their disease, but inevitably there will be some unexpected findings in the germline that have potential relevance to close relatives or may become relevant much later in life. Variation at over 40 loci is associated with the risk of developing type 2 diabetes, but lifestyle factors contribute far more to the absolute risk of developing the disease. Is it useful to feed back genetic information to the patient and their family when changing their lifestyle might have a greater effect?


Clinical translation from DNA sequence

Finally, the biggest challenges will be translation of our newly-gleaned knowledge of the DNA sequence into better clinical outcomes. Despite the early promises, genomic research has not yet created more personalised healthcare. This could be mainly due to the fact that genomics is still the new kid on the block, and these things take time to develop. New clinical trial designs will be required, which will be complex and will inevitably involve triaging the patient population to identify those with the right biomarkers for specific molecularly targeted agents. The lack of a properly informed workforce could be another stumbling block, something that is being addressed by HEE’s programme in genomic education. It might also be that there is quite simply more to life (and disease) than DNA alone. The DNA after all is just the start – identical twins only show some identical aspects. The way they lead their lives is often strikingly different which may result in very different health profiles.

As a final word – we are on a journey. All knowledge is useful, but we should  be aware of the risks and the benefits, both individually, and to society, and be cautious of overpromising to patients. It is better to use the genomic revolution to add to what we have already, rather than to throw the baby out with the bathwater. Pathology, as a discipline, needs to embrace the genomic revolution – if only to make sure we do not reinvent a more expensive wheel!  


Gerry Thomas is Professor of Molecular Pathology at Imperial College London’s Department of Surgery and Cancer.

* The views expressed are the opinions of the author and do not represent those of Imperial College London or the West London Genome Medicine Centre


Picture credit | Getty

Related Articles

The big question: molecular diagnostics

“Do you think the more we edge towards molecular diagnostics, the higher the likelihood that we will see biomedical science disciplines phased out in the future?”

Lab-on-a-chip for PCR

A new lab-on-a-chip infection test could provide cheaper, faster portable diagnostics, it is claimed.

National Genomic Research Library grows to over 110,000

The National Genomic Research Library (NGRL) now contains one of the richest genomic datasets in the world for both rare disease and cancer research.

NHS to pilot potentially revolutionary blood test

A blood test that may spot more than 50 types of cancer will be piloted by the NHS in a world-leading programme.