While the UK is making progress with its 100,000 genome project, finding answers to the question on how to translate genetic information into clinical benefit becomes more urgent. Do we need a 100,000 phenotype project?
[Valletta, MALTA] Personalised precision medicine is about tailoring medical care to the individual genetic make-up, or so proponents of large scale genomic sequencing want to make us believe. During eHealthWeek 2017, Tom Fowler, Deputy Chief Scientist of Genomics England, gave an update on one of the signature projects of the precision medicine community, NHS England’s 100,000 genome project. Considerable progress has been made in recent weeks, Fowler said, with about 30,000 genes already sequenced now.
The project was made possible by falling rates for sequencing: “Whole genome sequencing is below $1,000 now, or that is at least what we pay.” For its money, Genomics England receives 21 petabytes of data. Fowler used a nice analogy: “One petabyte would take 2,000 years to play on an MP3 player.” Given the current speed, NHS England should be able to finish its sequencing within 2,000 days, or indeed considerably less.
Benefits for patients?
The question remains, though, how to translate huge amounts of genomic data into palpable benefits for patients or citizens. A partial answer can be given in the area of rare diseases. Genome sequencing can help to identify rare genetic causes of various medical conditions, and this in turn helps both the patient and – by screening of parents and offspring – the close relatives. Fowler had an example of a patient who took part in the 100,000 genome project and in whom a rare gene mutation was detected that caused kidney failure.
These will be exceptions, though. To be able to tailor prevention and therapy to the individual genetic makeup in multi-genetic diseases would have far more impact in terms of numbers of patients involved. But this is also much more difficult. Charles Gutteridge, CCIO at Barts Health in London, argued that in order to be able to make sense of individual genetic data, in many cases detailed information on its counterpart, phenotypic data, will be needed. 'Phenotypic data' is about how high the individual blood pressure is, whether there is inflammation somewhere, how many hamburgers a person consumes, or how active he or she is physically.
The problem is that, typically, phenotypic information is not readily available in digital form at the moment. “The phenotypic world lags a long way behind each SNP in the genomic databases,” said Gutteridge. What could help, he suggested, was a broader adoption of SNOMED CT. This would make phenotypic information more available and, in particular, computable. And it would help geneticists both in research and patient care because they could draw on much more precisely defined phenotypic information that can be linked with the individual genes.
According to Gutteridge, inconsistent representation of the clinical context is among the biggest barriers to broad-scale adoption of precision medicine. A consistent approach to the digital representation of clinical features is urgently required, he said.