Turning science into healthcare

Cancer is the UK’s single biggest killer. In the UK, more than 166,000 people die from cancer each year (2016-2018) with around 375,000 new cases reported every year. Because cancer is more likely to occur as people age, we expect the number of cancer cases to rise as people live longer.

Cancer begins because of changes in genes within what was a normal cell. Although a cancer starts with the same DNA as the patient, it develops mutations or changes which enable the tumour to grow and spread. By taking DNA from the tumour and DNA from the patient’s normal cells and comparing them, the precise changes are detected. Knowing and understanding them strongly indicates which treatments will be the most effective.

Genomics has already started to guide and inform doctors about the best treatment for individual patients, but we still have more to discover. Many more cancer types, including those for whom there is hardly any successful current treatments such as lung cancer could be helped by knowing which gene changes were important.

Although rare diseases are individually very uncommon, because there are between 5000 and 8000 of them, a surprisingly large number of people are affected in total – 3 million – or, put another way 1 in 17 (or between 6 and 7 percent) of the UK population.

At least 80% of rare diseases are genomic, with half of new cases found in children.

Knowledge of the whole genome sequence may identify the cause of some rare diseases and help point the way to new treatments for these devastating conditions – vital progress given that some rare diseases take two or more years just to identify. As most rare diseases are inherited, the genomes of the affected individual (usually a child) plus two of their closest blood relatives were included to pinpoint the cause of the condition.

The project focused on cancer and rare diseases because both are strongly linked to changes in the genome.

In all, it was anticipated that about 75,000 people would be involved. The numbers added up like this: 25,000 cancer patients had two genomes sequenced each, so about 50,000 genomes in total. Roughly 51,000 participants had one genome sequenced each for rare disease (17,000 rare disease patients and two blood relatives each).

There was an extraordinary response by patients and their families wanting to take part in this ground-breaking project.

Today, we have sequenced over 100,000 genomes from over 97,000 patients and their family members, totalling over 21 petabytes of data – 1 petabyte of music would take 2,000 years to play.

Genomics England invested in the latest, state of the art sequencing machines to sequence the 100,000 genomes in the project. Because it was the first time sequencing had been attempted at such a scale in the UK, it was assumed that sequencing would be the most difficult part of the project. Whilst it wasn’t without its challenges, thanks to the support of our partner Illumina it proved to be less difficult than we had anticipated.

Data was a major challenge on two fronts. The first step after sequencing is to compare the possibly millions of differences between the patient’s genome and a reference genome, a process called variant calling. The next hurdle – annotation – is to interpret the meaning and importance of those differences which are important. Some of the differences will just be natural harmless variations between individuals, but some will be damaging and almost certainly involved in the development of disease. Automating this process – creating the Genomics England pipeline – so that it took weeks rather than years, was very difficult, but in the end it was achieved.

The second big data challenge was that information about a person and details of their illness are needed for interpretation. It's a bit like being able to measure the amount of fiber someone consumes in a day but not being able to say whether it's the recommended amount for them. To do that, you'd need to know more about the person – were they a child or an adult, for instance?

It’s a bit like being able to measure the amount of haemoglobin in a sample of blood but not being able to say whether it is normal without knowing more about the person who donated it – were they a child or an adult for instance?

Getting data from the NHS in such a way that it all followed the same ‘rules’ (so you knew you were comparing apples with apples) was very challenging, but the NHS staff involved worked incredibly hard to make it happen.

Another data issue was its size. The raw data from just one genome is about 200GB, which would occupy most of the average laptop’s hard drive. This mountain of data needed to be sifted, analysed and presented in a way that was helpful to doctors, most of whom would not have specialist knowledge of gene changes.

At one point, the cancer programme had to be halted because it became clear that the usual methods used in the collection and analysis of cancer tissues, such as preserving them in formalin and then fixing it in parafin (FFPE) damaged DNA.

To combat this, we found other ways to preserve samples, which is where we decided to use fresh frozen tissue. Again, the NHS was magnificent in responding to this challenge by completely reconfiguring how samples were collected.

With such a large amount of precious genome data, we addressed the challenge of keeping it secure by storing it with rigorous conditions for access controlled by a wide range of security measures as well as detailed governance. We also involved participants in deciding which researchers are allowed to access the data via the Participant Panel.

Each one of these challenges involved science at the cutting edge in a field that continues to move very rapidly, and doing it all at a scale never seen before.

Genomics England had to be very flexible, changing its plans frequently to reflect new advances but also being humble enough to learn from things that didn’t go right, especially in the pilot stage. We learned a huge amount from patients and clinicians that we continue to use to this day.