AI in the medicinal world seems to be really paying off! The researchers from MIT revealed that they used AI to discover an antibiotic compound goes by the name ‘halicin’ that can effectively kill even the highly resistant bacteria of several forms.
While different antibiotics existing present are not so effective in wiping out bacteria, halicin does so by destroying their capability for maintaining the necessary electrochemical gradient to yield energy-storing molecules. This way, it becomes tough for bacteria to endure. As such, E. coli when treated with halicin did not develop any sort of resistance within 30 days. On the other hand, when it was treated with ciprofloxacin, a conventional antibiotic, the bacteria fought it off within 3 days.
How AI Helped the Researchers?
Furthermore, the researcher’s team successfully developed such a system to discover molecular structures having certain desired traits, for instance, killing bacteria with more effectiveness as compared to past systems. On the contrary to previous systems and methods, the current systems with neural networks can learn molecular representations automatically and then map them in the form of continuous vectors to predict their traits and behaviours.
Once the mapping and vectors are ready, the team of researchers then trained their Artificial Intelligence (AI) on about 2500 molecules including both established and neutral products. In fact, there were a total of 1700 conventional or established drugs while natural products were a total of 800. AI found halicin to be the highest effective drug after skimming through a library of about 6000 compounds.
How Far the Research Has Gone?
However, the drug halicin might not be available in your prescriptions any time sooner. It is because the testing of the antibiotic is still on the run. As a matter of fact, MIT researchers have used the drug successfully in mice to eradicate an infection, A. baumanii, found commonly in US soldiers in Iraq and Afghanistan. But, they have not used this drug in humans yet.
Moreover, it is anticipated that a new trend to utilize AI in the medical world can start sooner. In fact, the scientists already have used the model in screening more than 100 million different molecules in some other database and found 23 candidates as well. In addition, they also aim to design drugs from scratch using technology and also to modify currently existing drugs in order to enhance their overall effectiveness while reducing their side effects. Thus, many lives could be saved by finishing off bacteria by using technologies like AI.