The neural network was taught to predict side effects from combinations of new drugs

The authors of the new work used machine learning to better predict side effects.

resulting from new drug combinations.

They collected data from reports of unwantedevents published by the U.S. Food and Drug Administration (FAERS). There are more than 15 million records in the databases. Further, the authors grouped events that often occur simultaneously to simplify the analysis and strengthen the relationship between the drug and its side effect profile.

The researchers then trained a neural network thatmimics the way the human brain makes associations between data so it can identify common patterns between drugs and their side effects.

To test their model, the researchersloaded data with obviously undesirable combinations to see how the neural network would react to them. The results showed that the model could recognize these new patterns and easily read the side effects of combination therapy.

We were able to identify individualtherapeutic effects using simple algebraic calculations. Because the algorithm is trained to recognize global patterns, it can accurately capture the side effects of combination therapy.

Bart Westerman, PhD, senior author of the study and associate professor at the Amsterdam Cancer Center

Read more:

After ten years of work, scientists questioned the standard model of physics

There is another “planet” inside the Earth: how it saved nascent life

Old cells were reprogrammed, and they became 30 years younger: how it works