AI has just helped discover a potent new antibacterial compound, arming the fight against antibiotic resistance.
Little would have Alexander Fleming thought his legacy would, at some point of time, be Artificial Intelligence’s (AI’s) to claim! But, AI has discovered a compound that could parallel Fleming’s discovery of penicillin. Scientists at the Massachusetts Institute of Technology discovered an antibacterial that beats one of the deadliest epidemics—drug resistance—with the help of AI. The AI programme was trained to recognise molecules that killed bacteria, being fed data on molecular and atomic features of 2,500 lab-made and natural compounds. It then analysed more than 6,000 compounds, focussing on compounds whose molecular/atomic particulars made them effective against the bacteria, but were unlike existing antibiotics.
In a matter of hours, it came up with halicin—named after HAL, the intelligent machine in 2001: A Space Odyssey—which was found to be effective against drug-resistant Acinetobacter baumannii, Enterobacteriaceae (two of the three pathogens the WHO lists as critical for new antibiotic development) and some others in mice tests. Scientists later made the algorithm work on 107 million compounds, and in three days, the programme reported 23 new antibacterials of which, the researchers say, two are particularly potent. As the global climate crisis releases new pathologies—with melting ice caps, ancient germs will be released—AI harnessed for pharmacological research can prove a boon. More so, since it is now being used for far more accurate diagnoses than was possible in the past.