Peptide2Mol: A Diffusion Model for Generating Small Molecules as Peptide Mimics for Targeted Protein Binding

2025-11-07

AI Summary

Existing AI-driven structure-based drug design often neglects endogenous protein–peptide interactions, leading to suboptimal peptidomimetic design. To address this, we propose the first E(3)-equivariant graph neural network diffusion model tailored for peptidomimetic generation. Our method jointly encodes the native peptide binding conformation and the 3D structural geometry of the target protein binding pocket, enabling binding-state-aware partial denoising and controllable molecular generation. The model adopts a non-autoregressive architecture and is trained on large-scale protein–ligand complex structures. Experimental results demonstrate that generated molecules significantly outperform state-of-the-art methods in peptide similarity (PepSim), predicted target binding affinity, and drug-likeness. Moreover, the framework provides interpretable, structure-guided lead discovery for protein–protein interaction targets—establishing a novel paradigm for rational peptidomimetic design.

Technology Category

• Machine Learning: Deep Neural Architectures and Foundation Models
• Computer Vision: Diffusion Models for Vision
• Natural Language Processing: Generation

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Economics, Online Markets and Human Computation: Uses of LLMs and GenAI for marketplace design, bidding, and strategic interactions

Abstract

Structure-based drug design has seen significant advancements with the integration of artificial intelligence (AI), particularly in the generation of hit and lead compounds. However, most AI-driven approaches neglect the importance of endogenous protein interactions with peptides, which may result in suboptimal molecule designs. In this work, we present Peptide2Mol, an E(3)-equivariant graph neural network diffusion model that generates small molecules by referencing both the original peptide binders and their surrounding protein pocket environments. Trained on large datasets and leveraging sophisticated modeling techniques, Peptide2Mol not only achieves state-of-the-art performance in non-autoregressive generative tasks, but also produces molecules with similarity to the original peptide binder. Additionally, the model allows for molecule optimization and peptidomimetic design through a partial diffusion process. Our results highlight Peptide2Mol as an effective deep generative model for generating and optimizing bioactive small molecules from protein binding pockets.

Problem

• Generating small molecules mimicking peptide-protein interactions
• Overcoming limitations of AI methods ignoring endogenous peptide binders
• Designing peptidomimetics using protein pocket and peptide references

Innovation

• Uses diffusion model for small molecule generation
• References peptide binders and protein pockets
• Enables molecule optimization via partial diffusion