Artificial intelligence has helped scientists find a “most powerful antibiotic” ever

Tram Ho

Artificial intelligence (AI) has helped scientists find a powerful antibiotic compound that can be used to kill the most dangerous viruses in the world today.

The team, led by Regina Barzilay and James Collins from Massachusetts Institute of Technology (MIT), used a machine learning algorithm to scan and screen a chemical compound database.

Their goal is to find the top candidates for a new antibacterial antibiotic, with the formula as different from existing antibiotics as possible.

According to research results published in the journal Cell, Barzilay and Collins have found an antibiotic compound and tested it on mice. This compound successfully kills bacteria that are resistant to all current antibiotics.

According to the researchers, this is the first time artificial intelligence – specifically machine learning algorithms that can cultivate their own capabilities – has been successfully applied to find new antibiotics.

“We want to develop a platform that allows to harness the power of artificial intelligence to usher in a new era in the discovery of antibiotics,” researcher Keith Collins at the Institute for Science and MIT’s Medical Technology (IMES) said in a statement.

Our strategy has resulted in the discovery of this amazing molecule. It can be said to be one of the most powerful antibiotics that humans have ever discovered.”

Trí tuệ nhân tạo vừa giúp các nhà khoa học tìm ra một loại kháng sinh mạnh nhất từ trước đến nay - Ảnh 1.

Artificial intelligence has helped scientists find a “most powerful antibiotic” ever

Antibiotic resistance is a phenomenon that occurs when bacteria are resistant to the antibiotics that humans use, after a while evolving to adapt to them. The World Health Organization (WHO) now proclaims that antibiotic resistance has now become one of the global threats, a situation requiring all governments around the world to take action. in many sectors.

Estimates show that more than 700,000 people die from antibiotic-resistant bacteria every year worldwide. If we do not offer strong solutions right now, the United Nations says the number could rise to 10 million by 2050.

This makes finding a new antibiotic a crucial task. In the past few years, the trend of antibiotic research has focused on only a limited number of targets, including some newly discovered antibiotics that have the same formula with existing antibiotics. resistant to bacteria.

Entering new wilderness areas, screening and hunting a whole new spectrum of chemical compounds, different from existing antibiotics, has yet to be conducted. And that’s where MIT scientists find their mission, with one powerful weapon they have: artificial intelligence.

We use artificial intelligence to create virtual molecular models and then predict their antimicrobial properties ,” Barzilay said. ” Typically, such refinements must be done in a lab, if so it is very slow and expensive. Now, a machine [with artificial intelligence] can display hundreds of millions of combinations. substance, in order to identify several candidates for testing. “

The low cost of this strategy allows us to explore a huge space of compounds, while testing only compounds with obvious potential. This is the first time AI has been used to identify identify a new antibiotic molecule , Barzilay said.

Trí tuệ nhân tạo vừa giúp các nhà khoa học tìm ra một loại kháng sinh mạnh nhất từ trước đến nay - Ảnh 2.

Artificial intelligence predicted halicin to be an effective antibiotic, and the truth is.

To do this, researchers first had to train their machine learning algorithms, to determine what are the characteristics of a chemical that allows it to wipe out E.coli bacteria. After the algorithm has been trained like that, the new team put it scanned through a database of about 6,000 pharmaceutical chemicals.

In this hunt, the algorithm identified a gravitational compound they named ” halicin “. Based on its properties, AI has predicted this to be an effective antibiotic, and it will work on principles that are different from existing antibiotics. Additional analysis shows that halicin does not poison cells.

The next step after finding the candidate is to test it. Researchers at MIT grew the bacteria in agar plates and found that halicin killed them. This allows them to advance to a test model on mice.

The mice were injected with a strain of A. baumannii, which is resistant to all antibiotics that humans had previously. As a result, halicin once again showed its effectiveness. The compound wiped out A. baumannii and helped the mice recover after just one day.

Not only that, but the team says that because it’s a completely new compound, far from existing antibiotics, halicin is also hard to resist. Tests showed that E. coli animals exposed to halicin continuously for a month did not evolve to resist it.

Trí tuệ nhân tạo vừa giúp các nhà khoa học tìm ra một loại kháng sinh mạnh nhất từ trước đến nay - Ảnh 3.

Halicin proved effective with ciprofloxacin, an old generation of antibiotics.

Even more good news, halicin is probably not the only potential candidate. In the future, we may have more new antibiotics, if we are willing to go into the wild, where there are unexplored chemical compounds.

Using its algorithm to scan through ZINC15, a huge database containing about 1.5 billion chemical compounds, MIT scientists continue to find 23 other candidates with similar antibacterial potential. as halicin.

Laboratory tests show that eight of the 23 compounds may also become new antibiotics. In the near future, they will continue to be tested on mice.

In addition to finding new antibiotic compounds, the researchers hope their work could also allow the design of stronger antibiotics, thanks to the enhanced properties of existing drug compounds. used current.

Commenting on the new research, Roy Kishony, a professor of biology and computer science at the Israeli Institute of Technology, said:

“This groundbreaking work has shown a change in the antibiotic detection model. This approach allows the use of machine learning algorithms at all stages of antibiotic development, from discovery to integration. new substance, to enhance its effectiveness and toxicity [with bacteria] “.

Refer to Feelitshareit

Share the news now

Source : Genk