ECNet is an evolutionary context-integrated deep learning framework for protein engineering

TLDR

Machine learning has been increasingly used for protein engineering, yet existing algorithms are limited because they capture general sequence contexts that are not specific to the target protein. The study introduces ECNet, a deep‑learning algorithm that exploits evolutionary contexts to predict functional fitness for protein engineering. ECNet integrates local evolutionary context from homologous sequences that explicitly model residue‑residue epistasis with global evolutionary context that encodes rich semantic and structural features from the vast protein sequence universe. ECNet accurately maps sequence to function, generalizes from low‑ to high‑order mutants, outperforms existing algorithms on ~50 deep mutational scanning and random mutagenesis datasets, and successfully guided engineering of TEM‑1 β‑lactamase variants with improved ampicillin resistance.

Abstract

Abstract Machine learning has been increasingly used for protein engineering. However, because the general sequence contexts they capture are not specific to the protein being engineered, the accuracy of existing machine learning algorithms is rather limited. Here, we report ECNet (evolutionary context-integrated neural network), a deep-learning algorithm that exploits evolutionary contexts to predict functional fitness for protein engineering. This algorithm integrates local evolutionary context from homologous sequences that explicitly model residue-residue epistasis for the protein of interest with the global evolutionary context that encodes rich semantic and structural features from the enormous protein sequence universe. As such, it enables accurate mapping from sequence to function and provides generalization from low-order mutants to higher-order mutants. We show that ECNet predicts the sequence-function relationship more accurately as compared to existing machine learning algorithms by using ~50 deep mutational scanning and random mutagenesis datasets. Moreover, we used ECNet to guide the engineering of TEM-1 β-lactamase and identified variants with improved ampicillin resistance with high success rates.