New AI popEVE: How Animal Evolution Helps Diagnose Genetic Diseases

Scientists from Barcelona and Harvard Medical School have introduced a new artificial intelligence model called popEVE, which can predict the onset of diseases due to genetic mutations. The researchers used data on the evolution of hundreds of thousands of animal species to build this model, which has proven to be more accurate than other methods, including the well-known AlphaMissense tool from Google DeepMind, according to the FT publication.

The popEVE model focuses on so-called missense mutations, where one amino acid in a protein changes. It analyzes these changes by comparing them with evolutionary information and identifies which variants are potentially harmful to humans.

During testing on data from 31 thousand families with children who had serious developmental disorders, popEVE correctly identified the most dangerous mutation in 98% of cases. In addition, the model identified 123 genes related to brain proteins that had not been previously considered in the context of similar diseases.

This new revolutionary model could have a significant impact on medical practice, especially in resource-limited countries, due to its energy efficiency and portability. Scientists are already successfully using popEVE in Senegal for diagnosing and treating patients, as well as for genome analysis in cases where there is no genetic data from parents.

Exciting developments are constantly occurring in the field of science. For example, scientists from the University of Cambridge have created a new device that mimics photosynthesis and safely converts carbon dioxide and water into useful chemical products without using toxic components and catalysts.