AI vs Antibiotic Resistance

A team at MIT have reported success in using machine learning (AI) to screen a variety of pharmaceutical molecules for their potential to kill antibiotic resistant bacteria (https://www.theguardian.com/society/antibiotics). Antibiotic resistance occurs following the overuse of antibiotics (in medical treatment and in farming, where it is used as a growth enhancer for meat production). Basically, the bacteria that better survive the application are the ones (with their resistant characteristics, of which there are a variety) going on to generate the new populations of bugs (it's pure Darwinian selection). These antibiotic resistant bacteria are particularly found in hospitals, nurseries and gyms and authorities have worried that their spread will return us to a pre-antibiotic era (with increasing death rates and enormous financial losses). The new method, basically fed into the programmes the characteristics of existing molecules (from drug trials or natural products) that might counter antibiotic resistance. In an initial trial, a failed drug for treating diabetes was found to be a likely candidate. In studies, this drug, now called Halicin (after the computer in the film 2001: A Space Odyssey), was found to be highly effective in treating a number of difficult antibiotic resistant conditions. Further trials are reportedly throwing up other candidate drugs much more quickly and cheaply than could be done by traditional means.