There are as many as 7,000 rare diseases — 95 per cent of which don’t have a cure. Together, they affect more than 400 million people worldwide — larger than the combined population of France, Germany and the UK.
Healx, a Cambridge-based company exhibiting at this week's Web Summit in Lisbon, is using Artificial Intelligence (AI) to predict treatments for rare diseases on a large scale.
Their work began with the story of Nick Sireau, whose children were diagnosed with Black Bone Disease. Sireau was told by doctors that there was no hope of treatment, but he didn’t take no for an answer.
He went on to create a patient group to help save his sons and worked tirelessly, but the process was too slow and too expensive, a problem many patient groups still face today. Luckily for Nick, they found a drug, originally developed as a weed killer, that offered possible hope. The drug is now on clinical trials, his children are on it and it has been a success.
The founders of Healx, Tim Guilliams and David Brown, had been advocating the use of AI for drug repurposing, and Nick became their inspiration.
“Our mission is really simple,” says Guilliams. "We want 100 rare disease treatments ready for clinical use by 2025, at a much lower price and much faster pace.”
The sweet spot of Healx is simple, Guilliams explains: “We use AI and Machine Learning to look at existing drugs to repurpose combinations for rare diseases."
Why use existing drugs?
“The best way of discovering a new drug is to start with an old drug," Guilliams says. "What is promising about AI and Machine Learning is that you can do this at scale. We have algorithms that go through the seven thousand rare diseases: select the best ones based on data, a number of criteria, and algorithms to match the treatments."
He continued: “If you combine two or three treatments together, you have about 13 billion possibilities per disease, so how do you select the best 10 or the best 20? This a Machine Learning problem, and we’ve been very happy because we’ve had remarkable success.
"One of our programmes after 18 months was ready to go through clinical trial for a fraction of the cost.”
A rare treatment opportunity
Beyond developing a successful business model, Healx is committed to delivering real change to rare disease patients with the creation of a “Rare Treatment Accelerator”. The programme seeks to provide a space to collaborate with selected patient groups to discover new treatments for rare disease patients. You can apply here.
Pharma is changing for good
“40 years ago, the advent of molecular biology and manipulation of the DNA led to the first biotechs and gave the first entrance to a whole new field that revolutionised the industry," Andrew Hopkins, CEO and Founder of Exscientia, during Web Summit.
“Today, machine intelligence is being applied not just in drug discovery, but also in patient population selection, clinical trial design and now in the advent of digital therapeutics. AI could be as big and as impactful on pharma as molecular biology was 40 years ago”, he said.
“A drug itself is a precision-engineered piece of technology every atom in it counts towards its properties.
Exscientia is a pioneer company in pharma, bringing together the power of AI and Machine Learning.
“We were really the first company to show how drugs can be designed and optimised using algorithms, we’ve now shown that we can design compounds far faster and far more cost-effectively using AI, rather than actually using the human endeavour," Hopkins said.
He said Exscientia had an “incredible year."
"The pharma industry is incredibly interested in our promise of how AI can deliver in terms of increasing productivity and also bringing more complicated molecules to the market and the clinic," he said.
He continued: “Our algorithms are agnostic to different diseases and agnostic to the type of proteins and drug targets which we can design against. If we’re going to try to revolutionise the pharma industry we can’t just be a niche player, we have to think about how we can disrupt the whole industry.
“However, when we develop for our own pipeline, we focus on three areas mainly: immune-oncology, immunology and oncology," he said.
Hopkins said AI was proven in drug design, and now "the next question is how we scale this across the industry.”