New Antibiotic From Artificial Intelligence: Kills Dangerous Bacteria

aNumak & Company
6 min readMar 8, 2022


Thanks to Artificial Intelligence (AI), antibiotics have been produced for bacteria that have developed drug resistance. Molecules that kill bacteria were taught to Artificial Intelligence (AI) to find an antibiotic that fights the world’s most dangerous bacteria.

It is one of the severe problems of our time that bacteria begin to develop resistance to existing antibiotics. The CDC (Center for Disease Control and Prevention) identifies antibiotic resistance as one of the most severe public health problems of our time. This is because the bacteria exposed to the antibiotic start to develop resistance against it and transfer this resistance mechanism to other bacteria. Therefore, this risk should be considered before using antibiotics. In addition, the unnecessary and incorrect use of antibiotics contributes to bacterial resistance and causes existing antibiotics to become dysfunctional over time. For example, the CDC (Center for Disease Control and Prevention) reports that too many antibiotics are prescribed unnecessarily in the USA. According to CDC (Center for Disease Control and Prevention) figures, approximately 47 million antibiotics are prescribed yearly for illnesses that do not require antibiotics, such as flu and colds. Therefore, we see that 30% of the antibiotics prescribed are unnecessary in these cases.

As early as 1945, Alexander Fleming foresaw this danger: “The public will demand [the drug]…then the era of abuses will begin.” Epidemiological studies indeed show that the overuse of antibiotics triggers resistance in bacteria. In bacteria, genes can be inherited from relatives and unrelated species, thanks to mobile genetic factors such as plasmids. This horizontal gene transfer allows antibiotic resistance to spread among different bacterial species. Resistance can also develop spontaneously, thanks to mutation.

The 1950s and 1960s were the golden age of antibiotic discovery. At that time, many new classes of antibiotics were discovered in rapid succession. But since then, very few new types of antibiotics have been found, and antibiotic resistance is slowly becoming a big problem.

Unless new, more potent antibiotics are developed, there may be a real possibility that minor injuries and common infections could kill people and that complex procedures such as chemotherapy and surgery could become too risky. The World Health Organization also states that antibiotic resistance poses a danger to the whole world, and if we are not careful, it can lead to massive epidemics.

Use of Antibiotics in Farm Animals

In addition to the unnecessary prescription of antibiotics, antibiotics for agricultural purposes also contribute to this crisis. All over the world, antibiotics are given to farm animals. Approximately 80% of the antibiotics sold in the USA are used in animals to support their development and prevent them from getting sick. Unfortunately, these antibiotics are given to animals also passed on to humans who consume the animals. The transmission of antibiotic–resistant bacteria from farm animals to humans was first reported 35 years ago. It was determined that bacteria resistant to antibiotics were found at high rates in the intestinal flora of farm animals and farmers. Recently, molecular detection methods have also shown that antibiotic–resistant bacteria in farm animals are passed on to humans who eat these animals.

Few New Antibiotics

The pharmaceutical industry could cope with this problem by producing new antibiotics in the past. Still, today, the production of new antibiotics has become more difficult due to economic and regulatory issues. As a result, 15 of the 18 largest pharmaceutical companies stopped producing antibiotics. Antibiotic research in academia also declined after cutting funding due to the economic crisis. In addition, antibiotic production is no longer seen as an economically profitable investment for pharmaceutical companies. Because they are used to treat disease quickly, they are not as good for companies as drugs used to treat chronic diseases.

For these reasons, a recent antibiotic discovery made by MIT researchers is of great importance. Researchers use artificial intelligence; to be more precise, they identified a new antibiotic using Machine Learning (ML).

Artificial Intelligence (AI) May Have Discovered a New Antibiotic

The algorithm applies the Machine Learning (ML) Method. This type of Artificial Intelligence (AI) analyzes large datasets to find patterns and access certain information. He then uses the data he analyzes to make new predictions. In this case, the algorithm analyzed millions of molecules and chemicals to find the component Halicin that could be used as an antibiotic. In an interview with Wilbur, Collins says, “We asked the model which of the molecules in our database that were unlike the existing antibiotics in our database could be anti–bacterial.”

Collins thinks this algorithm can also be used in cancer, neurological diseases, and metabolic disorders. However, the methods used by Deep Learning algorithms are not specific to antibiotics.

Bioengineer James Collins said the drug was “one of the most powerful drugs ever found,” adding: “It is remarkably effective against a wide range of antibiotic–resistant pathogens.” Antibiotic resistance occurs when bacteria evolve and become resistant to drugs.

Scientists have announced that they have found an antibiotic that will kill a type of bacteria among the most dangerous bacteria globally and have developed resistance to all drugs using Artificial Intelligence (AI).

The drug in question was found to feed an Artificial Intelligence (AI) with a different structure from existing antibiotics and a learning feature with a digital library of compounds used in pharmacy.

In tests, it was found that the drug killed the bacteria acinetobacter Baumann and Enterobacteriaceae, two of the three pathogens in urgent need of medicine identified by the World Health Organization.

Regina Barzilay, the senior researcher from the research team, explained that this is the first time that a drug has been created in this way.

Researchers taught Artificial Intelligence (AI) the molecules that kill bacteria to find the new antibiotic. For this, approximately 2,500 drugs and natural compounds were used.

After the algorithm understood what should be in an effective antibiotic, the scientists also introduced 6,000 building blocks that were likely to be used in drugs.

From that point on, the algorithm began digital experiments with building blocks different from existing antibiotics but considered adequate.

This allowed the current antibiotic to be very different from its predecessors to kill resistant bacteria.

Researchers named the drug Halicin after the robot Hal in 2001: A Space Journey.

It was determined that Halicin also killed the bacteria that caused tuberculosis.

Announcing their discovery in the journal Cell, the researchers stated that among the building blocks of the drug is a molecule that was tried to be used against diabetes but was later abandoned.

After this discovery, scientists introduced 1.5 billion building blocks to Artificial Intelligence (AI) instead of 6,000 building blocks and enabled them to experiment with over 107 million of them.

After three days of computation on the computer, the program returned with 23 potential antibiotic recommendations. Two of them turned out to be highly effective.

Researchers announced that they want to continue their audits on this issue.

Jonathan Stokes, one of the authors of the study, explained that it is impossible to obtain 107 million building blocks in traditional ways and to conduct experiments one by one in laboratories. Still, they can try it digitally with Artificial Intelligence (AI):

“Being able to do these experiments on a computer is a huge time and cost advantage.”

Researcher Barzilay said that they plan to conduct experiments in the future for antibiotics that kill bacteria more selectively and thus do not harm healthy bacteria in the gut.

Click for more information on this topic;



aNumak & Company

aNumak & Company is a Global Business and Management Consulting firm with expertise in building scalable business models for diverse industry verticals.