With the cost of drug discovery rising, more pharmaceutical companies are trialling ways of using the latest artificial intelligence (AI) systems to improve their chances of bringing a new drug to market.
Due to their impressive data-processing power, AI systems can be used to identify re-purposing opportunities more quickly and efficiently than would be possible using traditional scientific research methods. For example, an AI system could be used to analyse existing research data in order to establish whether a drug molecule will bind to other specific targets. Unlike research activities undertaken by humans, this form of analysis is guaranteed to be objective as it is based on patterns derived from known data sources.
In order to extend the usefulness of AI systems, some pharmaceutical companies are working together to improve the quality of datasets and use them to train algorithms through a process of machine learning. A recent initiative known as the Melloddy project, involving a number of big pharma companies, is using a blockchain system to store research data on a secure ledger.
Among the early drug discovery success stories is Atomwise, which has been using deep neural networks to analyse simulations of molecules in order to reduce the time required to synthesise and test compounds. Its proprietary technology, known as AtomNet, has since been used to find a treatment for Ebola virus infections that had caused the death of over 11,000 people in Africa and some other parts of the world and Merck has also been using it to identify opportunities to re-purpose existing medicines.
As their use becomes more widespread, AI systems could help to find a cure for debilitating medical conditions, potentially affecting large numbers of people, where it has proved difficult to find a cure by other means – such as Alzheimer’s. They could also prove useful in finding treatments for rare conditions affecting relatively few people, that might otherwise struggle to attract funding.
Digitisation and the increased availability of diverse patient health data – everything from genetic information to sensor data from wearable devices – is encouraging the use of AI systems in precision medicine. AI-based diagnostic tools can analyse and process this data to design treatments for individual patients – for example, an AI-based eye-imaging tool is currently used by clinicians to diagnose diabetic retinopathy.
Of course, as is the case in any fast-moving area of R&D, it is important for those involved in developing these AI systems and using them for drug discovery should take advice on how to protect their inventions. This protection will ensure they can commercialise their inventions fully, without risk of a third party copying what they have done.
Despite the promise of AI systems in drug discovery, some hurdles remain which could undermine their capabilities. First and foremost, regulatory authorities may be cautious about approving new drugs identified by algorithms and AI-based techniques. However, pharmaceutical companies have much to gain from finding a way to harness the power of AI and machine learning, using it to identify and bring new drugs to market. It makes sense for companies to act now and protect their intellectual property as they go.
By Dr Joanna Thurston, partner and patent attorney at European intellectual property firm, Withers & Rogers.