How AI helps drug research

The opportunities and dangers of artificial intelligence (AI) are discussed intensively, with the positive possibilities of AI increasingly fading into the background.

This is due to the many warnings about the dangers of the machines – especially from people who are well versed in them: the CEOs of OpenAI, the company behind Chat-GPT. Together with a group of scientists, the AI ​​icons addressed the public with a short but drastic appeal:

In doing so, they poured even more oil into the fire of the further heated discussion. However, the latest scientific findings show that AI is not only a danger to humans:

It usually takes years to bring a new drug to market. As a new study from Ohio State University shows, AI can significantly speed up this drug development process.

In the case study, researchers examined whether AI is capable of developing synthetic routes of existing drugs, including mavacamten, a drug used to treat systemic heart failure, and oteseconazole, used to treat fungal infections.

The result: AI was able to develop not only the patented drugs, but also other alternatives – and at a rapid pace. The study was recently published in the journal Nature.

Study leader Xia Ning writes that AI not only saves time and money in drug development, but may also offer better properties than naturally occurring molecules. And further:

Scientists from McMaster University in Canada and the Massachusetts Institute of Technology have already succeeded in developing an antibiotic using AI. The research team focused on one bacterium: Actinetobacter baumanii. A pathogen that is mainly feared as a hospital bacteria. The germ sticks to surfaces such as door handles and can cause diseases such as pneumonia, meningitis or wound infections.

But maybe not for long.

In animal tests on mice, the antibiotic was able to prevent infection. AI came into play in the search for active substances: the developed machine examined 7500 chemical substances against the germ. There was actually one active substance among the substances sorted out by the AI. In this way, the machine has saved man a lot of work. Before the drug can be marketed, it must first be tested in human clinical trials. (cst)