Protein Directed Evolution

1 Product Overview

This product focuses on optimizing ligand proteins to significantly enhance their binding affinity and stability with receptor proteins, while maintaining the overall conformation unchanged. Using cutting-edge computational design tools, especially ProteinMPNN, we ensure the integrity of protein function and structure during the optimization of either the full-length protein or specific key regions. The service is structure-centric, combining deep learning-driven sequence optimization tools with high-precision binding energy analysis methods. By targeting the key interaction sites of the ligand protein, we can quickly generate optimization solutions.

This product is widely applicable in fields such as drug target protein optimization, industrial enzyme design, antibody and antibody fragment optimization, and molecular recognition in bioengineering. The service offers significant advantages in efficiency, accuracy, and customization, greatly shortening experimental validation cycles, reducing research costs, and providing tailored optimization solutions to clients. With the support of rapid iteration design and screening, we help clients obtain proteins with high affinity and stability, meeting the needs of various research and industrial applications.

2 Product Screening Workflow

    2.1 Identification of Modification Regions

We precisely construct the ligand-receptor protein complex model and analyze the key interaction sites and functional domains of both proteins. Based on this analysis, we determine the regions of the ligand protein, either full-length or specific parts, to target for optimization.

    2.2 ProteinMPNN Iterative Design

Using the optimized complex model as a basis, we employ the ProteinMPNN tool for sequence design of the ligand protein. While maintaining the overall conformation and key structural domains (such as functional residues), we optimize other sites by adjusting the amino acid composition and output the candidate ligand protein for evaluation.

    2.3 HDOCK Binding Energy Evaluation

Each round of modified candidate proteins is evaluated for binding energy, quantifying the stability of their interaction with the receptor protein. The best binding affinity designs are selected based on energy scoring (the lower the binding energy, the more stable the binding).

    2.4 AlphaFold Overall Evaluation

We use AlphaFold to predict the structure of the optimized candidate protein-receptor complex and analyze its overall confidence and composite score. Compared to the original ligand-receptor complex results, if the optimization score is lower than the reference, the design will re-enter the ProteinMPNN iterative process for further modification and evolution. If the score is higher than the reference, the new ligand protein sequence is finalized.

3 Deliverables