Artificial intelligence (AI) has been used to address antibiotic resistance. An antibiotic that’s able to kill superbugs has been discovered with the help of a machine-learning algorithm.
Scientists have revealed that use of AI has helped in producing an antibiotic compound, halicin that cannot only kill many forms of resistant bacteria but also kill them in a new way. Where many antibiotics are slight spins on existing medicine, halicin can be used to kill bacteria by affecting their ability to maintain the electrochemical gradient necessary to produce energy-storing molecules. It is difficult for the bacteria to develop then: in E. coli, resistance to the drug was not seen in 30 days; it could, however, prove resistant to the conventional antibiotic cipofloxacin within three days.
The potent antibiotic is structurally not like any other antibiotic seen before. Researchers from Massachusetts Institute of Technology (MIT) made use of a novel computer algorithm to sift through a vast digital archive of over 100 million chemical compounds that could be set up. They were able to spot those that were able to kill bacteria using different mechanisms from existing drugs. This method highlighted a molecule that appeared to possess some truly remarkable antibiotic properties. The team named the molecule halicin (from the AI system, termed ‘Hal’ featured in the movie by Stanley Kubrick, called 2001: A Space Odyssey). This was reported in the journal Cell.
The team has developed a system that could find molecular structures with desired traits (say, killing bacteria) more effectively than before. Unlike what was done earlier, now the neural networks were configured with representations of molecules automatically, mapping them into continuous vectors that help predict their behaviour. Once ready, the researchers trained their AI on 2,500 molecules that included both 1,700 established drugs and 800 natural products. With these, they came up with 6,000 compounds, and AI was used to conclude that halicin would be highly effective.
In mice, halicin was able to effectively treat tuberculosis and drug-resistant Enterobacteriaceae (a kind of bacteria including E. coli and Salmonella). It was also very effective against Clostridium difficile, a ‘stomach bug’ that often sweeps through hospitals, and another drug-resistant bacterium that can cause infections of the blood, urinary tract, and lungs. It was used to eradicate A. baumanii, a common infection, in mice, but it hasn’t been used in human trials.
The team wants to design antibiotics from scratch and modify existing drugs to increase their effectiveness or reduce their unintended side effects.
Post-halacin
Following the success of halicin, the team returned to the database and used the AI algorithm to identify 23 compounds that were structurally dissimilar to existing antibiotics and non-toxic to human cells. Some 8 molecules could have antibacterial properties, with 2 being very potent. They could be used for tackling superbugs and antibiotic-resistant infections.
Due to the overuse of antibiotics, some potentially dangerous bacteria have evolved drug resistance, which means the diseases cannot be treated. This is a trend that may continue as more and more bacteria gain resistance to conventional drugs.