Rotamer-free protein sequence design based on deep learning and self-consistency

Several previously proposed deep learning methods to design amino acid sequences that autonomously fold into a given protein backbone yielded promising results in computational tests but did not outperform conventional energy function-based wet experiments. Here we present the ABACUS-R method, which uses an encoder–decoder network trained using multitask strategy predict sidechain type of central residue from its three-dimensional local environment, includes, besides other features, types conformations surrounding sidechains. This eliminates need reconstruct and optimize structures, drastically simplifies sequence process. Thus iteratively applying different residues is able produce self-consistent overall for target backbone. Results experiments, including five structures solved by X-ray crystallography, show outperforms state-of-the-art success rate precision. A method on backbones, ABACUS-R, this study. shows improved performance when compared with