A method for multiple-sequence-alignment-free protein structure prediction using a protein language model

Abstract

Protein structure prediction pipelines based on artificial intelligence, such as AlphaFold2, have achieved near-experimental accuracy. These advanced pipelines mainly rely on multiple sequence alignments (MSAs) as inputs to learn the co-evolution information from the homologous sequences. Nonetheless, searching MSAs from protein databases is time consuming, usually taking tens of minutes. Consequently, we attempt to explore the limits of fast protein structure prediction by using only primary structures of proteins. Our proposed method, HelixFold-Single, combines a large-scale protein language model with the superior geometric learning capability of AlphaFold2. HelixFold-Single first pre-trains a large-scale protein language model with thousands of millions of primary structures utilizing the self-supervised learning paradigm, which will be used as an alternative to MSAs for learning the co-evolution information. Then, by combining the pre-trained protein language model and the essential components of AlphaFold2, we obtain an end-to-end differentiable model to predict the three-dimensional coordinates of atoms from only the primary structure. HelixFold-Single is validated on datasets CASP14 and CAMEO, achieving competitive accuracy with the MSA-based methods on targets with large homologous families. Furthermore, HelixFold-Single consumes much less time than the mainstream pipelines for protein structure prediction, demonstrating its potential in tasks requiring many predictions. AlphaFold2 has revolutionized bioinformatics, but its ability to predict protein structures with high accuracy comes at the price of a costly database search for multiple sequence alignments. Fang and colleagues pre-train a large-scale protein language model and use it in conjunction with AlphaFold2 as a fully trainable and efficient model for structure prediction.