Deep learning-based prediction of protein structure using learned representations of multiple sequence alignments

Deep learning-based prediction of protein structure usually begins by constructing a multiple sequence alignment (MSA) containing homologues of the target protein. The most successful approaches combine large feature sets derived from MSAs, and considerable computational effort is spent deriving these input features. We present a method that greatly reduces the amount of preprocessing required for a target MSA, while producing main chain coordinates as a direct output of a deep neural network. The network makes use of just three recurrent networks and a stack of residual convolutional layers, making the predictor very fast to run, and easy to install and use. Our approach constructs a directly learned representation of the sequences in an MSA, starting from a one-hot encoding of the sequences. When supplemented with an approximate precision matrix, the learned representation can be used to produce structural models of comparable or greater accuracy as compared to our original DMPfold method, while requiring less than a second to produce a typical model. This level of accuracy and speed allows very large-scale 3-D modelling of proteins on minimal hardware, and we demonstrate that by producing models for over 1.3 million uncharacterized regions of proteins extracted from the BFD sequence clusters. After constructing an initial set of approximate models, we select a confident subset of over 30,000 models for further refinement and analysis, revealing putative novel protein folds. We also provide updated models for over 5,000 Pfam families studied in the original DMPfold paper. Significance StatementWe present a deep learning-based predictor of protein tertiary structure that uses only a multiple sequence alignment (MSA) as input. To date, most emphasis has been on the accuracy of such deep learning methods, but here we show that accurate structure prediction is also possible in very short timeframes (a few hundred milliseconds). In our method, the backbone coordinates of the target protein are output directly from the neural network, which makes the predictor extremely fast. As a demonstration, we generated over 1.3 million models of uncharacterised proteins in the BFD, a large sequence database including many metagenomic sequences. Our results showcase the utility of ultrafast and accurate tertiary structure prediction in rapidly exploring the "dark space" of proteins.