Newborn genome programs
Clinicians and anyone who has had a baby in the UK will be familiar with the newborn blood spot test, which screens infants for nine rare conditions that benefit from early treatment. In mid-2023, the NHS and Genomic England hope to take this to the next level, by performing a pilot study of WGS of newborn babies to expand testing to many more rare diseases. If the pilot is successful, WGS for newborns will be implemented into routine NHS care.6
The aim of the study is very focused—the NHS will analyze genome sequences only for conditions that emerge during childhood and for which clinical interventions exist, rather than returning every possible bit of information, says Simon Wilde, engagement director at Genomics England. Guidelines such as these are the result of a careful program of public engagement performed in the light of Genomic England’s experience with the 100,000 Genomes Project. Public acceptance is key for the success of the programme, says Wilde.
The team is also consulting rare disease experts to make sure that all the conditions to be reported have treatment pathways and care packages available. “It’s about making sure that we’ve got everything in place before we begin.”
The results of the first public dialogue exercise showed that the public are supportive of the technology, as long as certain conditions and safeguards are met. These include data privacy and informed consent. “That’s what we’ve been able to take forward and at least begin to start thinking about: what would a program of sequencing genomes in babies look like?” says Wilde.
The team is also studying factors such as the best time during pregnancy to approach women with information about the programme.7 “We need to make sure that the offer we make is one that’s acceptable, that’s understandable, and that works for parents,” says savages.
The project will begin in summer 2023 and aims to sequence up to 100,000 newborn genomes in the first instance, with the possibility of going beyond that if needed.8
Genomics England has recently launched a research programme, in collaboration with the NHS, called Cancer 2.0. “This is not yet clinical, but it has clinical intent,” says Moss. Cancer 2.0 has two main aims: to explore new DNA sequencing technologies and to unite imaging data with genomic data, to develop new insights into cancer and improve diagnosis and treatment.
The new sequencing technologies center on nanopore sequencing, which can read long sections of DNA and so detect changes involving large chunks of the genome that are hard to spot with current “short read” technologies. This is important because “structural variations increasingly look like they are druggable directly,” says Moss. It can simultaneously detect DNA methylation patterns that are key to understanding epigenetic contributions to a cancer.
The imaging arm of the Cancer 2.0 project involves integrating radiology and digital pathology images with genome data from the 15,000 cancer patients in the 100,000 Genomes Project. In partnership with Leeds Teaching Hospitals NHS Trust and the National Pathology Imaging Co-operative, the project will create a large dataset of more than 250 000 digital pathology images, alongside detailed diagnostic information from pathology reports.9
Combining the molecular features of the tumor with spatial images should give you much more specificity about whether a patient will respond to treatment, says Moss. “Response to treatment to drugs is not always simply dependent on the molecular biology of the tumour, but also about the microenvironment of the tumour,” he explains. Applying machine learning to the data should also yield deeper insights into the disease and anticancer treatments.10 It might ultimately be possible to use these algorithms to diagnose cancers in the clinic.
How well a drug works, the risk of adverse side effects, and what the optimal dose is can vary considerably between patients. Much of this variability is affected by genetics, and researchers have identified variants affecting patient response to more than 40 drugs. Testing for these variants, known as pharmacogenomics, is still limited within the NHS. But moves are underway to expand the range of tests available, and even to take a proactive approach that involves testing patients before they need treatment and recording the results for future reference, says Munir Pirmohamed, a clinical pharmacologist, and David Weatherall, chair of medicine at the University of Liverpool.
A small number of genetic tests for adverse drug reactions are already available on the NHS, including HLAB57 testing for hypersensitivity to the anti-HIV drug abacavir, and dihydropyrimidine dehydrogenase (DPD) deficiency to prevent serious toxicity from a class of anticancer drugs known as fluoropyrimidines .11 This identifies patients who should receive lower doses of the drugs to minimize adverse side effects.
But there is a clear need to expand these tests to cover more drugs.12 What’s more, an increasing number of drug manufacturers are stipulating genetic testing on their medicines to guide dosing. Some of these tests are not available on the NHS and must be done privately. “It is important we think about how we can make sure to make that available because people will develop adverse effects that we could have prevented,” says Pirmohamed. Being better able to predict drug efficacy will help clinicians avoid trial and error when treating some conditions like depression and select the best drugs faster. “Even if it increases the predictability a little bit, then we may reduce some suffering for our patients,” says Pirmohamed.
Pirmohamed’s own work shows just how dramatic this reduction can be. In 2004, his team showed that genetic testing for abacavir hypersensitivity was cost effective.13 By 2006, all NHS HIV clinics were using the test and hypersensitivity rates dropped from 5%-7% to less than 1%.
Instead of single gene tests, Pirmohamed and his colleagues are advocating panel tests, where patients are typed for a range of variants and these are kept on the patient’s electronic health record. “You’ve already got the data available rather than having to do it again,” he says. His team is working with Our Future Health, a forthcoming research program that aims to recruit five million UK residents. Pirmohamed’s team has designed the gene chip that will be used to type participants for pharmacogenomic variants.14 Ethical discussions are still underway, but it is possible that the data will be returned to individual participants.
Another project to which Pirmohamed is contributing, UP-Gx,15 recently ran the Preemptive Pharmacogenomic Testing for Preventing Adverse Drug Reactions trial,16 which recruited 6900 people from across Europe, including patients from the Royal Liverpool University Hospital. The study pre-emptively tested participants with a panel of 40 markers in 13 genes associated with drug responses. Doctors used this information when prescribing drugs for patients in the test group (those in the control group received standard care) with the aim of reducing adverse drug reactions. The data from the study are now under review at a medical journal.
The field is not without obstacles: like many other genomics projects, phamacogenomics suffers from a lack of ethnic diversity in its study subjects. For example, the NHS currently tests for four variants in the DPD gene, but these variants are all derived from European ancestry populations, says Pirmohamed. There are no tests for rare DPD variants in patients with African backgrounds, meaning they might wrongly be classed as wild-type. Pirmohamed is currently working on a project in the UK and internationally to identify more of these variants in a range of ethnicities. He also has a project underway to develop genetic testing for warfarin dosing in populations in sub-Saharan Africa.
Ultimately, a patient’s whole genome sequence could be kept on record and then interrogated for phamacogenomic data, says Pirmohamed. But patients need to be involved in the design of these services. “They need to understand what the advantages, but also the limitations, are,” says Pirmohamed. Delivering pharmacogenomics also means creating clinical decision support systems so that busy primary and secondary care clinicians don’t have to spend a lot of time trying to decipher genetic test results, he adds. Another matter is applying pharmacogenomics to people taking multiple drugs, a particular problem in elderly patients.
Pirmohamed sees pharmacogenomics as an evolution, rather than a revolution, in clinical practice. “The way I look at it, the predictivity that we have at the moment is pretty low,” says Pirmohamed. “If we can achieve an increase in predictivity for particular drugs to more than 50%, more than 60%, then we’re doing a much better job than we are at the moment.”
Data security and confidentiality are sensitive areas when it comes to gaining and keeping public confidence in technologies such as genomic medicine.
Seeking, gaining, and maintaining consent are key, says Wilde. Ongoing public engagement and developing best practices for data management is “a foundational part of what we do,” he says, and has been since the inception of the 100,000 Genomes Project. “It’s something that we take very seriously. It’s something that we know is a live issue and will continue to be a live issue.”
Genomics England works on a model where it obtains consent before any data are stored and ensures that people understand what they might be used for, how anonymous they might be, and what security safeguards are in place. “What we’d like to do is try and talk about what the potential risks and benefits might be, but ultimately make sure that people have that choice about whether to take part in the first place,” says Wilde. “But if they ever get cold feet or, for whatever reason, they don’t want to take it anymore, that there are mechanisms in place for them to have their data removed and destroyed.”