202110220937 Protein folding

Protein folding is the process of determining the 3D shape of a protein from its amino-acid sequence. The 3D shape determines the protein's function and linking the amino acid sequence to the structure therefore links the sequence to the function. This has long been regarded as one of if not the hardest problem known to humanity due in part to 202110221009 Levinthal's paradox.

DeepMind's Breakthrough

DeepMind's program for determining the 3D shapes of proteins stands to transform biology¹

The Google owned AI company DeepMind made a massive leap in 2020 in our ability to predict how protein folding works. DeepMind's AlphaFold outperformed all the competitors by a wide range in the biennial Critical Assessment of Structure Prediction (CASP) contest. One scientist even went so far as to say, "in some sense the problem is solved."

It's a game changer. This will change medicine. It will change research. It will change bioengineering. It will change everything.

— Andrei Lupas, Max Planck Intitute for Development in Biology¹

In many cases, the structure predicted by DeepMind was indistinguishable from the state of the art experimental methods of directly measuring protein shapes.

Though much of DeepMind's predictions were so close that it was unclear whether their model or the experimental measure was more accurate, they did struggle to model individual structures in protein complexes.

Impacts of Predictive Methods

• Mis-folded proteins are responsible for several neurodegenerative diseases including Alzheimer's, ALS, and Parkinson's. With better understanding of the folding process, is it possible for us to synthesize better drugs or proteins to combat these diseases?
• #thread