‘Robot scientist’ could speed up search for new drugs

An artificially-intelligent ‘robot scientist’ called Eve could pave the way in the bid to make drug discovery faster and cheaper, according to the researchers who developed her.

Eve, who is based at the University of Manchester, can test thousands of compounds every day and uses machine learning to continually fine-tune her approach. She recently discovered that a compound known to have anti-cancer properties could potentially be used to combat malaria.

There has been a significant increase in the use of automated machinery in science over the past few years and robot scientists represent the latest development in workplace artificial intelligence.

They are able to develop and test hypotheses to explain any observations that they make, perform experiments and then interpret the results to amend their hypotheses accordingly. This cycle is then repeated. The experiments are designed and performed automatically by computer, without the need for human input, meaning that robot scientists are ideal for recording scientific knowledge.

Eve is the younger sister of Adam, a robot scientist prototype who was developed at the University of Cambridge in 2009 with the help of researchers from the University of Aberystwyth. He became the first machine to independently discover new scientific knowledge when he autonomously investigated the genomics of the baker’s yeast, Saccharomyces cerevisiae.

The same team developed Eve and based on her recent endeavours, she promises to be even better. She has been focusing on identifying drug candidates for malaria, as well as other tropical diseases such as African sleeping sickness and Chagas’ disease.

Professor Ross King of the Manchester Institute of Biotechnology said, “Neglected tropical diseases are a scourge of humanity, infecting hundreds of millions of people, and killing millions of people every year.

“We know what causes these diseases and that we can, in theory, attack the parasites that cause them using small molecule drugs. But the cost and speed of drug discovery and the economic return make them unattractive to the pharmaceutical industry.”

That’s what makes Eve so useful. She is much more efficient when it comes to multitasking, saving valuable time and money. Professor Steve Oliver of the University of Cambridge explained, “In the case of drug screening, she can search her library and select compounds that have a high probability of being active against the chosen drug target and she will prioritise screening them.”

Eve works by screening a large set of compounds against assays (tests), which are designed to be automatically engineered. This means that more types of assay can be applied and improves the efficiency of process. The probability of making an important discovery is subsequently increased.

Eve is capable of screening more than 10,000 compounds a day. Despite the simplicity of this process, it is still relatively slow and wasteful. The system is also unintelligent, since it makes no use of what is learnt during the procedure.

To improve this, a random subset of the library is selected and tested against the first assay. By taking note of which compounds pass, Eve uses statistics and machine learning to predict new structures that might perform even better against the assays. It is hoped that future robot scientists could even synthesise these compounds.

The researchers tested this approach by developing assays targeting key molecules from the parasites responsible for various tropical diseases. These assays were tested against roughly 1,500 clinically approved compounds.

Eve discovered that an anti-cancer drug also inhibits DHFR, a key molecule in the malaria parasite. Given the fact that many strains of parasite are developing immunity to existing drugs, this discovery could prove to be extremely important.

“Despite extensive efforts, no one has been able to find a new antimalarial that targets DHFR and is able to pass clinical trials,” concluded Professor Oliver. “Eve’s discovery could be even more significant than just demonstrating a new approach to drug discovery.”