Halicin: The AI Made Superbug Killer
Can Artificial Intelligence discover an antibiotic effective against super-resistant bacteria to the formulas currently on the market? The answer is yes, and it also has a name: Halicin.
A few weeks ago the publication, in the magazine Cell, on the identification of a particularly powerful pharmacological formula against a series of superbugs. The research was conducted by specialists from the renowned MIT, one of the most important research universities in the world based in Cambridge, Massachussets, in the United States. The study underlines that an algorithm was the one to discover the antibiotic, thus officially marking the entry — on a global scale — of machine learning into the scientific field for the research of new drugs.
The term Halicin is inspired by Stanley Kubrik’s famous masterpiece, “2001 A Space Odyssey”, and the result obtained from the antibiotic developed by the MIT research team is equally noteworthy: the formula would have managed to wipe out dozens bacteria, including some that until then had proved resistant to drugs already on the market and were therefore on the World Health Organization’s list of “special guards”.
According to Regina Barzilay, Professor of Computer Science at MIT: “In terms of the discovery of unpublished antibiotics, what happened represents an absolute novelty.”
James Collins, professor of Bioengineering at the same institute, is also of the same opinion: “I think it is one of the most powerful antibiotics ever discovered in history, because it shows a remarkable activity in contrasting a very wide range of antibiotic-resistant pathogens.”
This last comment is linked to the fact that not even the dreaded E. Coli bacterium is able to resist the action of the new antibiotic, as demonstrated by the algorithm in less than a month of work.
The experiment: how artificial intelligence identified the antibiotic Halicin
The algorithm that discovered the Halicin antibiotic was “trained” to recognize the molecular structure of about 2500 formulas. More or less half of them corresponded to drugs approved by the FDA, Food and Drugs Administration, the US government body that regulates pharmaceutical and food products, responding directly to the directives of the US Department of Health and Human Services, while eight hundred were found in nature.
The researchers programmed the system to track not only molecules with antibiotic properties, but to specifically focus on those that exhibited different structures than those typical of existing antibiotics. Using a second machine learning program, the scientists then examined the results by identifying compounds that are safe for humans.
It is also important to note that, in addition to Halicin, eight other molecules deemed potentially promising and therefore of interest for further tests would have been discovered.
According to the first results, Halicin attacks the cell membranes of bacteria, effectively affecting their ability to produce energy. To protect itself from the action of the antibiotic, the bacterium would in practice have to undergo two or more genetic mutations — a difficult result to achieve even for particularly resistant superbugs.
In terms of experimentation, the new antibiotic was tested on a series of bacterial colonies and, subsequently, also on mice. The result? Halicin was able to completely eliminate infections caused by a bacterial strain that until then had proved resistant to all other known formulas.
The research team now plans to continue their studies with the collaboration of a pharmacological company or, alternatively, of a non-profit organization, with the ultimate aim of creating a truly effective and truly safe drug for use. on man.
Artificial Intelligence and Medicine: effective synergy in a crucial historical moment
For Professor Collins, such a noteworthy result could not have come at a more appropriate time: “We are facing a growing crisis regarding antibiotic resistance, and this situation has arisen due to both an increase in pathogenic microorganisms become resistant, both due to an increasingly scarce production of new antibiotics by the biotech and pharmaceutical industries. “
Halicin’s discovery confirms that the technological evolution of Artificial Intelligence makes these “super-smart” algorithms increasingly part of our daily life, not only in areas such as home automation, marketing or industry, but also in medicine: from the timely diagnosis of pathologies that are difficult to identify up, as in this case, to the discovery of formulas and drugs for the treatment of particularly complex bacteria and viruses.