Application of AI and ML in the discovery of new drugs: A new way to fight diseases

In recent decades, researchers and pharmaceutical companies have made enormous efforts to find new drugs and preparations to fight various diseases, including cancer and neurodegenerative diseases. Artificial intelligence (AI) and machine learning (ML) have recently begun to play a key role in this quest.

In an interview with NEWS.am Tech, intern researcher Philipp Seidl of the Johannes Kepler University in Linz says that AI and ML can significantly speed up the process of discovering new molecules, which can later become the basis of new drugs.

Molecular modeling

In the search for potential drugs, scientists turn to molecular modeling methods. It is a computational method used in chemistry to model the behavior and properties of molecules in drug discovery. This method involves creating computer images of molecules and their interactions to understand their structure, properties and possible biological activities.

Scientists use molecular modeling to predict how molecules might interact with specific biological targets. This method is also used to create new molecules with desirable properties, such as those that can inhibit the growth of cancer cells or penetrate the blood-brain barrier.

Today, one of the most urgent problems for scientists in this field concerns the approaches to represent molecules in simulations. Scientists have not yet reached a consensus over this issue: some believe that molecules can be represented as a graph, others as a three-dimensional object, perhaps moving.

Quality data

In order to search for new molecules that can become the basis for new drugs, a large amount of high-quality data is needed, based on which AI will learn. Many specialists, however, claim that such data are currently insufficient, especially in the field of biology and chemistry.

As Philipp Seidl mentioned, scientists can find the data they need if they want. Many biologists work with large numbers of molecules and share their research results on the Internet. In addition, pharmaceutical companies have a large database. Over time, the number of high-quality datasets on which AI algorithms will be based will only increase.

Speeding up the process and difficulties along the way

In the past, it took many years to find and test new drugs. However, with the use of new methods, including ML and AI, this process, as Seidl noted, can be cut at least in half.

Today, specialists are already conducting clinical trials of drugs developed using AI and ML. ML models have long been used to narrow the selection from a large number of potential molecules in the early stages of drug development.

The main difficulty here is that scientists still do not understand the mechanisms by which many diseases develop, and this makes it difficult to find molecules that can be used in the treatment of these diseases.

"For example, should we look for molecules that affect serotonin receptors in depression? There are many disputes about this now," said the specialist.

The use of AI and ML in the discovery of new drugs is a huge advance in science. However, there are still many challenges in this field, including a lack of high-quality data on which to train AI models, and the difficulties in understanding biological processes and diseases. Despite this, every year modern technologies make discoveries in the field of medicine more accessible and faster.