Researchers at Manchester University have devised a way to predict the best combinations of drugs to treat specific illnesses from billions of possibilities.
The process, which involved applying a complex computer algorithm to a robotic testing machine, could be used to develop new drug combinations to combat a wide variety of diseases and conditions.
‘What you want to do is find a cocktail of things you can add but you can’t do millions of experiments,’ research leader Prof Douglas Kell told The Engineer.
‘So what you do is have a set of candidate solutions, look how those do and based on the results of that find another set of candidate solutions so you evolve a cocktail.’
The researchers developed the process by testing drugs for treating inflammation, which can develop after a stroke and cause further damage to the body and even death.
Using a cocktail of drugs can be more effective than just one but it also allows patients to take smaller doses, reducing potential toxicology concerns.
Each stage in the process required trialling 50 combinations in experiments using a robotic laboratory device that would pipette the drugs into an assay and detect when cells were inhibited from producing a certain chemical.
Once the experiments were complete, the evolutionary computer algorithm used the results to select the next set of drug combinations to trial and the process would begin again. After 10 stages the algorithm recommended the best combination.
The robotic equipment is commercially available and the researchers have published the algorithm so that the system may be repeated by anyone who wants to trial drug mixtures in this way. Each stage can be completed in as little time as one day.