Evolutionary-scale prediction of atomic-level protein structure with a language model

🤖 Plain-English Summary

Recent advances in machine learning have leveraged evolutionary information in multiple sequence alignments to predict protein structure. This results in an order-of-magnitude acceleration of high-resolution structure prediction, which enables large-scale structural characterization of metagenomic proteins.

🔑 Key Findings

• We demonstrate direct inference of full atomic-level protein structure from primary sequence using a large language model.
• As language models of protein sequences are scaled up to 15 billion parameters, an atomic-resolution picture of protein structure emerges in the learned representations.
• This results in an order-of-magnitude acceleration of high-resolution structure prediction, which enables large-scale structural characterization of metagenomic proteins.

🔗 Related Resources