Protein Binder Design via Surface Fingerprint Segmentation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current methods for de novo protein-protein interaction design face challenges in generating amino acid sequences that form stable complexes with target proteins, particularly when no structural elements from preexisting binders are known, due to weak energetic signatures and the difficulty in finding compatible protein scaffolds for novel binders.

Innovation Solution

A geometric deep learning framework, MaSIF, is used to generate surface fingerprints that capture geometric and chemical features critical for protein interactions, enabling the design of novel protein binders by predicting target sites and identifying binding seeds that can engage these sites effectively.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If hotspot-centric approaches and rotamer information fields are used to design de novo PPIs, then the method provides a systematic framework for placing residues on target interface, but the weak energetic signatures from scoring functions to single side chain placements and difficulty in finding compatible protein scaffolds remain unresolved

Engineering Contradiction:

Improvesystematic framework for residue placementVSAvoidbinding affinity prediction accuracy

Core Design Contradiction:

Ease of manufactureVSReliability

Solution Approach 1:

The patent segments the protein interface into distinct regions: buried/core interface regions and interface rim regions. This segmentation allows different design strategies to be applied to each region - the buried regions provide strong binding affinity through hydrophobic interactions, while the rim regions provide specificity through polar interactions, thereby resolving the contradiction between systematic framework and binding reliability

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent applies different chemical and structural properties to different regions of the interface. The buried interface regions are designed with hydrophobic residues for strong binding, while the rim regions use polar residues for specificity. This local differentiation enables the system to achieve both reliable binding affinity and precise target recognition simultaneously

Inventive Principle:

Principle #3Local quality

2Adaptability or versatility

If de novo binders are designed for flat interfaces that lack deep pockets, then the method can target previously undruggable sites, but the weak energetic signatures from scoring functions make it difficult to achieve high affinity binding

Engineering Contradiction:

Improveability to target flat interfacesVSAvoidbinding affinity

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

The patent transitions from relying solely on deep pocket geometry (traditional 3D binding sites) to utilizing the 2D surface topology of flat interfaces. By analyzing surface fingerprints and chemical properties across the entire interface surface, the method identifies binding opportunities on flat regions that were previously considered undruggable, while maintaining reliable binding through optimized residue placement patterns

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

3Extent of automation

If no structural elements from preexisting binders are known, then the method can truly achieve de novo design, but the challenge of generating amino acid sequences that form stable complexes becomes significantly more difficult

Engineering Contradiction:

Improvede novo design capabilityVSAvoidcomplex stability

Core Design Contradiction:

Extent of automationVSReliability

Solution Approach 1:

The patent changes the parameters used for evaluating binding potential from traditional structure-based approaches to surface fingerprint-based chemical property analysis. By using descriptors that capture the chemical nature of interface regions rather than relying on preexisting structural templates, the method achieves automated de novo design while maintaining the ability to generate stable complexes through physics-based scoring functions

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS20230395187A1Systems and methods for de novo design of protein interactions with learned surface fingerprints

Publication Date: 2023.12.07 IMPERIAL COLLEGE INNVOATIONS LTD

AI summary

The present application relates to a computer-implemented systems and methods for protein interaction design using surface fingerprints. The method comprises predicting at least one target interface site with high binding propensity, wherein, optionally, the step of predicting at least one target buried interface site comprises generating at least one surface fingerprint associated with a protein interaction based on at least one protein interface, wherein the at least one surface fingerprint preferably embeds geometric and/or chemical features of molecular surfaces, identifying at least one binding seed that displays required features to engage the target site, and performing a binding seed transplantation to protein scaffolds to confer stability and additional contacts on the designed interface.