Exploring “dark-matter” protein folds using deep learning

De novo protein design explores uncharted sequence and structure space to generate novel proteins not sampled by evolution. A main challenge in de novo design involves crafting “designable” structural templates to guide the sequence searches toward adopting target structures. We present a convolutional variational autoencoder that learns patterns of protein structure, dubbed Genesis. We coupled Genesis with trRosetta to design sequences for a set of protein folds and found that Genesis is capable of reconstructing native-like distance and angle distributions for five native folds and three novel, the so-called “dark-matter” folds as a demonstration of generalizability. We used a high-throughput assay to characterize the stability of the designs through protease resistance, obtaining encouraging success rates for folded proteins. Genesis enables exploration of the protein fold space within minutes, unrestricted by protein topologies. Our approach addresses the backbone designability problem, showing that small neural networks can efficiently learn structural patterns in proteins. A record of this paper’s transparent peer review process is included in the supplemental information. •Genesis VAE efficiently learns to transform simple protein fold representations into designable models•Coupled with trRosetta, Genesis enables de novo design of proteins with specific architectures•Genesis-trRosetta rapidly explores tbe dark-matter protein fold space Harteveld et al. introduce Genesis, a convolutional variational autoencoder that explores untapped protein sequence and structure spaces. Genesis generates stable de novo proteins, including novel “dark-matter” folds, validated through high-throughput protease resistance assays, demonstrating its ability to efficiently expand the protein fold space.