Friday, 29 July 2022 10:22

AI network predicts structure of 200 million proteins from 1 million species Featured

As Nature writes in its current issue, researchers have used the artificial intelligence network called AlphaFold to predict the structures of about 200 million proteins from 1 million species, covering almost all known proteins on the planet.


Most drugs are developed based on structural information. However, developers today still rely predominantly on methods such as X-ray cristicalallography analysis for this purpose. AlphaFold produces highly accurate predictions of the 3D shape or structure of proteins and also information about the accuracy of the predictions. In the process, 35% of the predictions were reported to be very accurate, and 45% were still reliable enough for many applications.

AlphaFold was developed by London-based company DeepMind, a Google company, based on Deep Learning.

The data will be freely available in a database in the future.

Further information:
Callaway, E. (2022). 'The entire protein universe': AI predicts shape of nearly every known protein. Nature News, 28.07.22. https://doi.org/10.1038/d41586-022-02083-2