How Genomics England is going beyond genomics, and beyond England — The BioCentury Show

CEO Chris Wigley on Genomics England’s next steps, and how it’s addressing near-term hurdles faced by the whole field

As it approaches its 10-year anniversary, Genomics England can count a few achievements in the past decade, which largely serve as a foundation for the ambitions it has for the next one. Those goals include expanding to other omics, improving the diversity of the database and, broadly, establishing itself as an engine of growth for data in the U.K.

The goal to expand is part of the tenet to continually enrich the dataset: Genomics England cares about more than DNA and more than just England, said CEO Chris Wigley, speaking on The BioCentury Show.

Founded in 2013 to commandeer the 100,000 Genomes Project, the organization’s early challenge was to achieve its goal at a time when the cost of sequencing a genome times 100,000 was far more than its budget. Sequencing costs came down, efficiencies went up, and in late 2018 the organization completed the sequence of its 100,000th genome. Wigley joined as CEO in 2019, coming from a background in machine learning and AI.

Work continues with the original participants, said Wigley.

He said Genomics England has built a foundation with NHS to offer whole genomic sequencing for rare disease and cancer, making England the first country to have nationwide whole genome sequencing for diagnostics.

The dataset is also established as a research resource for investigators in academia and industry.

Genomics England’s business model is an “aquarium model,” said Wigley. “You can come in and study the fish. You can’t take the fish away with you. We’re not a fish shop. We don’t sell data.” Instead, researchers can study the data stemming from patients with a specific disease or with a specific variant, and look at a patient level, at anonymized genomic and phenotypic representations. Researchers can also request via Genomics England to recruit patients for a trial, though that process is not yet optimally streamlined, said Wigley.

The organization is adding high definition pathology images, transcriptomics, proteomics, metabolomics and more — part of its response to users who articulated some of those needs. It’s also an example, according to Wigley, of how Genomics England tackles the challenge of becoming a collection of data for its own sake, and builds it as a tool to furnish and enable innovation.

Wigley said the investment the country had made in genomics underpinned the strength the U.K. showed in the pandemic in rapidly sequencing the virus and deploying that information, with clinical data from its integrated healthcare system, to advise treatment strategies. “The fact that we had invested in genomic sequencing in the healthcare system already meant that there were regional labs that had Illumina machines that were good to go,” said Wigley.

One broad concern across the omics field is that the existing databases are heavily skewed to genomes coming from individuals of white, European ancestry, which means that therapies and strategies based on information from those resources can not only be less relevant for patients from other ethnic origins, but can exacerbate health disparities further. 

“At a technical level this is very clearly a very real risk that we need to address,” said Wigley. This was something learned with the development and rollout of machine learning with facial recognition. “If you train the model on a bunch of photos of white men, guess what? It performs more poorly if you’re a person of color, if you’re a woman, if you are older, if you are younger, etc.” Overfitting of these models is a “real issue” he said.

With genomics, one challenge is that there is not a “causal relationship” between a box on ethnicity that someone self-identifies and their DNA. Those two things have a relationship but it’s not a direct and simple relationship, said Wigley.

That means gathering more data, first, plus expanding the framework of reference genomes. “We’re going from one reference genome that you’re assessing rareness against to about 40-ish now, and aiming to get to about 350,” said Wigley. Genomics England received $25 million funding to specifically address the problem around diverse data.

Wigley also talked about how the organization is engaging individuals over the longer term to provide longitudinal data. One example is a recently started pilot to sequence DNA from 100,000 babies and follow them through life.