Google DeepMind has announced the SynthID Bio system, designed to tag synthetic proteins that were designed using generative artificial intelligence.
This system is capable of adapting to various data types. It introduces minor yet detectable changes in the amino acid selection within peptide chains while simultaneously adjusting atomic coordinates according to the protein's three-dimensional structure.
Modeling and practical tests demonstrated the effectiveness of this system using several proteins: Vascular Endothelial Growth Factor A (VEGF-A), the SARS-CoV-2 coronavirus spike protein RBD, and the immune checkpoint PD-L1.
The results showed that the introduced modifications do not negatively affect the proteins' functionality, their ability to bind to ligands, or the natural diversity of the sequences, while remaining easily recognizable.