An AI-designed drug will enter human trials in ‘world-first’
In a world first, a drug molecule created by machine learning will be used in human trials in Japan.
The critical landmark for AI-driven drug discovery comes from a tie-up between UK startup Exscientia and Japanese pharmaceuticals firm Sumitomo Dainippon Pharma.
The drug, DSP-1181, was created in just 12 months, and will be used to treat patients who have obsessive-compulsive disorder (OCD). The process of drug development can typically take up to four and a half years.
The increasing use of AI in healthcare is laying the foundations for a huge market in years to come. A report from healthcare analytics firm Definitive Healthcare found one-third of hospitals and imaging centers to now be using AI and machine learning technologies to accelerate the task of patient diagnosis and treatment.
Drug discovery, in particular, is a focal application of AI: ‘Traditional’ methods of developing a treatment are estimated to cost a staggering $US2.6 billion, and much of this is evaporated in the roughly nine out of 10 candidate therapies that fail.
The use of AI can make drug discovery faster, cheaper and more effective for patients across a range of illnesses from cancers to heart disease. The life-saving and lucrative potential has led to a surge in companies developing AI drug discovery products. There are close to 200 startups currently using machine learning technology to research and develop drugs. Markets & Markets predicts the AI drug discovery industry will explode at a CAGR of 41 percent, hitting a worth of US$1,434 million by 2024, up from US$259 million last year.
Machine learning has already been successfully applied in expanding the number of patients who can benefit from existing medicines. But drug-design itself has been much more challenging.
“The design and development of molecules through medicinal chemistry has always been a slow and laborious process,” said Sir John Bell, the Regius Professor of Medicine at Oxford University, in Financial Times.
“Exscientia can do this in many fewer steps, which is really impressive, and it comes from very sound scientific principles. I think they are a real asset to have in the UK.”
Exscientia claims to leverage drug discovery data and experienced “drug hunters” to “generate candidates” in “roughly one-quarter the time of traditional approaches.”
According to Professor Andrew Hopkins, a molecular biologist and Exscientia’s CEO, the firm developed AI that learns faster than conventional approaches, allowing his team to test just 350 compounds – a fifth of the normal number of compound candidates “which is record-breaking productivity,” he said.
“The algorithms […] can be applied to any drug targets, against a huge range of diseases in oncology, cardiovascular and rare diseases.”
DSP-1181 will enter phase one trials in Japan, with global tests to follow if successful. The firm, which is now involved in projects with pharma giants Beyer and Sanofi, is also working on potential drugs for the treatment of cancer and cardiovascular diseases.
Commenting on the breakthrough, Prof. Hopkins said that 2020 will be the year of the first AI-designed drug, “but by the end of the decade all new drugs could potentially be created by AI.”