Researchers have developed synthetic, soluble versions of cell membrane proteins to facilitate faster and easier drug screening. These proteins are typically difficult to access and study due to their location in the cell membrane. By redesigning them as soluble analogues, the proteins can now be produced in bulk and directly interact with molecular targets in solution. This approach is more cost-effective and efficient compared to traditional methods that rely on observing cellular reactions or extracting membrane proteins from mammalian cells.
Using deep learning technology, the researchers were able to design these soluble protein analogues based on existing protein structures. The AI platform AlphaFold2 was used to predict amino acid sequences for these proteins, which were then optimized for functionality and solubility using another deep learning network called ProteinMPNN. This innovative approach successfully produced soluble proteins that retained their native functionality, even for complex protein structures like the G-protein coupled receptor (GPCR), a major pharmaceutical target.
The results of this study demonstrate the potential of this pipeline for vaccine research and cancer therapeutics. By making these proteins more accessible and easier to work with, researchers can accelerate the discovery of new drugs and antibodies. This research was published in the journal Nature, showcasing the impact of deep learning technology on protein design and drug discovery.
Source link
Source link: https://www.drugtargetreview.com/news/151158/the-potential-of-deep-learning-generating-drug-targets/