by Bennett Davis
Using a machine learning program, artificial intelligence software at MIT has designed a new antibiotic able to kill dozens of bacteria, including some that defeat all other known antibiotics.
The developers called the find “amazing” and believe it ranks among the most important antibiotics ever discovered because of its range of effectiveness.
The new drug is named Halicin; “-icin” is a common ending for antibiotic names and “Hal” salutes the computer in the film 2001: A Space Odyssey.
The researchers designed their computer model to look for chemical features that make molecules effective at killing E. coli bacteria. To do so, they showed the model about 2,500 compounds, including 1,700 FDA-approved drugs and a set of 800 natural remedies with diverse structures and a wide range of bioactivities.
The resulting machine learning project, which can screen more than a hundred million chemical compounds in a matter of days, was designed specifically to imagine potential antibiotics that can kill bacteria using different mechanisms than those of existing anti-bacterial drugs. Halicin was the result.
In tests, E. coli bacteria didn’t become resistant to Halicin within a 30-day treatment period in mice. In contrast, in no more than three days, the bacteria developed a defense against ciprofloxacin, a common antibiotic used to treat a range of infections.
Halicin also was found to do no damage to healthy tissue or to beneficial bacteria that live in the human digestive tract.
TRENDPOST: Antibiotic resistance is a high-priority health emergency as bugs that once succumbed to antibiotic treatments no longer do. Halicin is the first of a new generation of antibiotics designed and tested by computer that can defeat resistant bacteria and will replace older drugs as human clinical trials are completed.
AI programs such as this one also can be refined to custom-design therapeutic compounds for individuals based on their unique body chemistry.
