Google DeepMind, the artificial intelligence (AI) division of tech giant Google, has recently unveiled a groundbreaking tool called AlphaMissense. This tool has the ability to predict whether genetic mutations are likely to be harmful, which could have a significant impact on the research of rare diseases. AlphaMissense is specifically focused on missense mutations, which involve single-letter alterations in the genetic code.
In the human genome, there are approximately 9,000 missense mutations. Some of these mutations are harmless, while others can lead to diseases such as cystic fibrosis, cancer, or impact brain development. Out of the four million missense mutations observed in humans, only 2% have been classified as disease-causing or benign, leaving a staggering 71 million mutations unclassified.
However, AlphaMissense is capable of predicting 89% of these mutations with an impressive 90% accuracy rate. The tool assigns a score to each mutation, indicating its potential for causing disease or its pathogenicity. Through its predictions, AlphaMissense identified that 57% of the mutations were likely benign, 32% were probably pathogenic, and the remaining mutations remained uncertain.
Scientists can access the tool’s database, which is a valuable resource for further research. The results and research study accompanying the release of AlphaMissense were published in the well-regarded journal Science. In an article published in Science, experts Joseph Marsh and Sarah Teichmann praised AlphaMissense for its superior performance compared to existing tools.
While these predictions are not primarily intended for clinical diagnosis, Jun Cheng from Google DeepMind explained that they could help increase the rate of diagnosing rare diseases. Furthermore, the tool’s predictions could potentially lead to the discovery of new disease-causing genes, offering opportunities for new treatments.
AlphaMissense was trained on human and closely-related primate DNA, enabling it to recognize widespread genetic mutations. Cheng compared this training process to language learning, highlighting that the tool can process millions of protein sequences and understand regular protein sequences. This understanding facilitates the identification of mutations and their potential harm.
Overall, AlphaMissense has the potential to revolutionize the field of genetics and contribute to advancements in diagnosing rare diseases and discovering new disease-causing genes. Google DeepMind’s groundbreaking tool demonstrates remarkable accuracy and has garnered praise from experts for its superior performance. As the scientific community gains access to this resource, it is hoped that further breakthroughs will be made in understanding and treating genetic mutations.
