How UK artificial intelligence organisations are transforming the future of drug discovery

The United Kingdoms’ most innovative start-ups are tapping into new areas of drug discovery through the power of AI. Exploring the latest developments is Tom Bethell, business development director at Kao Data, the high performance data centre company which is sponsoring the Start-up of the Year award at the 2023 Cambridge Independent Science and Technology Awards.

Emerging technologies such as artificial intelligence (AI) and machine learning (ML) are helping to revolutionise the development of drugs, including treatments for Covid-19, cancer research and dementia.

Just as AI and ML have transformed entire sectors such as banking and financial services, so too are they transforming biomedical science. Behind the scenes, for example, smart algorithms are analysing vast datasets at scale and at almost instantaneous speeds. The superhuman power of AI has, therefore, put the technology at the very centre of the work being undertaken by today’s drug discovery pioneers - those such as Lifebit and Exscientia.

The role of AI in drug discovery was given a real sense of urgency by the Covid pandemic, and this urgency hasn't gone unnoticed by investors. In 2022, almost half of the UK's most generously-funded start-ups are in the life sciences and pharmaceutical sectors. Research bodies are teaming up with AI firms such as InstaDeep, which has recently gained the interest of pharmaceutical pioneer BioNTech, to achieve breakthroughs in drug development that wouldn't ever have been possible without the insights delivered by GPU-accelerated technologies.

In this article, we’ll take a closer look at these organisations, and reveal how they're using AI and ML to drive ever more fast, accurate, scientific discoveries that will save and change lives.

Using AI to develop cancer treatments

“In the future, all drugs will be designed with AI,” proclaims Oxford-based Exscientia – which is busy proving that prediction correct. Already the world's leading pharmatech, Exscientia has nearly half a billion dollars ($474.4m) of funding at the latest count, and has been using the funds to accelerate its application of AI-based techniques for small-molecule drug discovery. Most of its projects currently focus on oncology (cancer treatment), but Exscientia's work also covers research into Covid, inflammation and immunity (for example arthritis) and psychiatry.

As a “full-stack AI drug discovery company”, Exscientia deploys the latest AI techniques in a whole host of ways. The firm claims to be both the first to automate drug design and the first to have an AI-designed molecule enter clinical trials. It also puts ever-more sophisticated algorithms at the heart of drug discovery, applying its AI systems to develop novel, precision-engineered treatments that have a high chance of offering real therapeutic benefit.

AI is central to data analysis at Exscientia, too.

“AI is simply better at learning from complex data sets than humans, and better at determining what to do next,” says the company, adding that faster learning shortens discovery times and improves outcomes for patients and their families.

Developing Covid treatments with deep learning

Kao Data client InstaDeep has long used AI and deep learning to guide business decisions in sectors from robotics to finance. Previously nominated by CB Insights as one of the world's 100 most promising AI startups, UK-headquartered InstaDeep is now emerging as a key biomedical research partner, thanks in part to its work with pharmaceutical giant BioNTech.

BioNTech first selected InstaDeep as its AI partner in 2019. Together the companies began to develop novel immunotherapies for a range of cancers and infectious diseases, and more recently BioNTech has announced its intention to acquire InstaDeep to strengthen its pioneering position in the fields of AI-powered drug discovery, design, and development.

While BioNTech provided the vast internal and external datasets, InstaDeep – with the help of Kao Data's high performance infrastructure – provided the AI and ML algorithms to analyse it, delivering deep insights and accelerating the pre-clinical drug discovery stage.

Following extensive AI-powered research into the development of next-generation vaccines and biopharmaceuticals for Covid, both firms announced in January 2022 that they'd developed and successfully tested an AI-based early warning system to detect potential high-risk Covid variants. “Early flagging of potential high-risk variants could be an effective tool to alert researchers, vaccine developers, health authorities and policy makers,” said BioNTech founder and CEO Dr Ugur Sahin. “This provides more time to respond to new variants of concern.”

Providing the power needed for AI research

Using AI at the cutting edge of biomedical research requires serious compute power, low latency connectivity and customised infrastructure architectures. AI, for example, is extremely data intensive, so if a research body wants to incorporate AI or machine learning into its work, and it wants it to be fast, accurate and scalable, then it needs an appropriate GPU-accelerated computing environment to support it.

Since 2021, Kao Data has provided a bespoke high-performance computing (HPC) environment that optimises the learning process at the heart of InstaDeep's AI decision-making technology. This custom-built supercomputing environment is specifically configured to give the company, and its users, access to all compute nodes in the data storage cluster, enabling faster machine learning capabilities.

Not all AI start-ups and medical research bodies will need to own premium supercomputing resources, and we must remember that in the fields of high performance computing, one size doesn't fit all. NVIDIA’s Cambridge-1 supercomputer is a perfect example of ‘pooled HPC resources’ in action, providing the compute, the processing power, and the GPU-accelerated technology to find cures for humankind's most tenable diseases.

Using AI to accelerate gene-based drug discoveries

Researchers in genomics – the study of genes, a vital tool in drug discovery – play a bigger role in medicine development than ever, and more so today thanks to the power of artificial intelligence.

Organisations are using AI algorithms to identify disease-causing mutations, design therapeutic targets, identify treatments and ultimately cut the turnaround time to get new, effective drugs on the market. Little wonder that the ‘AI in genomics market’ is expected to grow at around 43% annually over the next seven years, reaching $6.22 billion by 2029.

UK-based Lifebit is at the forefront of the AI revolution in genomics. The company is building the world’s first AI-powered genomics platform that understands DNA data and generates meaningful insights with the intelligence of humans, but at much greater speed and scale.

Many other organisations are developing AI tools to accelerate gene-based pharmaceutical research. These include NVIDIA, whose Clara Healthcare Platform helps medical scientists develop treatments for Covid, and Congenica, whose AI analysis of complex genomic data supports research into treatments for rare diseases and cancer by the Hong Kong Genome Project, Sanford Health and the University of Glasgow.

Identifying treatments for heart disease and MND

UK firm BenevolentAI is the first fully-integrated AI company with pharmaceutical discovery and clinical development capabilities. However, it still puts collaboration with partners at the centre of its approach, teaming up with AstraZeneca, Novartis and the Universities of Southampton, Glasgow, Edinburgh and more to apply AI and ML to unlock biological insights and develop medical treatments.

With funding of $292m to date, BenevolentAI has used GPU-accelerated technologies to help find treatments for a whole range of conditions. To name just a few examples, its proprietary AI-enabled drug discovery platform, the Benevolent Platform, has identified candidates to treat the devastating disease amyotrophic lateral sclerosis (AMD, also known as motor neuron disease or MND), chronic kidney disease, and the fatal lung disease idiopathic pulmonary fibrosis (IPF).

In collaboration with AstraZeneca, BenevolentAI is also using machine learning to interrogate vast amounts of data and uncover high-quality potential treatments for IPF. The collaboration is set to continue for at least two more years to study heart failure and systemic lupus erythematosus.

What’s clear is that AI will undoubtedly be an integral component of future drug discoveries, but the infrastructure systems supporting them will also need to evolve in line with new technological developments including GPUs, supercomputers, and potentially, quantum computers. At Kao Data, our high performance data centres are precision engineered to support and enable the future of drugs discovery, and we believe that the platform will be vital for the next-generation of pharmaceutical start-ups seeking to harness the power of AI.