Dr Martin-Immanuel Bittner
Chief Executive Officer, Arctoris Ltd
Drug discovery starts in the lab, yet is slow and error prone. Combining robotics and artificial intelligence (AI) is the key to enable faster progress and bring new drugs to patients sooner.
All industries have been transformed by the introduction of electronics, computers, automation, machine-to-machine communication and the cloud. However, there is one area that has not kept pace with innovation: not much has changed in biomedical laboratories since the days of Alexander Fleming and his discovery of penicillin.
Researchers still hunch over lab benches carefully pipetting reagents by hand, recording data and taking notes on paper. Lab work, the key to discovering potential new drugs, is trapped in the past.
The consequence: R&D productivity in this area has decreased 15-fold in the past five decades. Bringing a new drug from the lab bench to patients now costs $2.6bn and often takes up to 15 years. The COVID-19 pandemic has made it clear that we don’t have this kind of time to spare.
To switch drug development R&D productivity from reverse gear into full speed we have to completely rethink the way we discover new drugs and embrace the benefits of automation and AI.
The rise of AI
The application of AI in drug discovery has already shown great promise. Last year, Hong Kong-based Insilico Medicine sped up the process of drug design using its AI approach, shortening this critical step from the pharma average of 1.8 years to just 45 days.
In April this year, UK’s BenevolentAI used machine learning to find that an approved Eli Lilly drug could be repurposed as a treatment for COVID-19. The drug went on to receive FDA approval for emergency use in COVID-19 patients in November.
Automating experiment execution and data capture generates the large, richly annotated data sets that are the key to drug discovery success, both for human researchers and for AI, thereby accelerating the path from the lab to the patient.
Robots to the rescue
Yet, this is only the beginning of the AI revolution in drug discovery. Only one challenge remains in the way: the key to unlocking the true potential of AI lies in the availability of large amounts of structured, standardised, well annotated data.
This has become the bottleneck, since manually collected data is riddled with errors and shortcomings. The proliferation of vague, unstructured data leads to inconsistencies and misleading results, estimated to cost $28bn annually in the US alone.
Automation is the answer: using robotics to conduct drug discovery experiments liberates scientists from manual labour, allows them to focus on tasks requiring human creativity and ingenuity, and enables new insights to be based on rich, reliable, meaningful data.
In summary, automating experiment execution and data capture generates the large, well annotated data sets that are critical for drug discovery success both for human researchers and for AI, thereby accelerating the path from the lab to the patient.