Abstract:Biologists frequently desire protein inhibitors for a variety of reasons, including use as research tools for understanding biological processes and application to societal problems in agriculture, healthcare, etc. Immunotherapy, for instance, relies on immune checkpoint inhibitors to block checkpoint proteins, preventing their binding with partner proteins and boosting immune cell function against abnormal cells. Inhibitor discovery has long been a tedious process, which in recent years has been accelerated by computational approaches. Advances in artificial intelligence now provide an opportunity to make inhibitor discovery smarter than ever before. While extensive research has been conducted on computer-aided inhibitor discovery, it has mainly focused on either sequence-to-structure mapping, reverse mapping, or bio-activity prediction, making it unrealistic for biologists to utilize such tools. Instead, our work proposes a new method of computer-assisted inhibitor discovery: de novo pocket-aware peptide structure and sequence generation network. Our approach consists of two sequential diffusion models for end-to-end structure generation and sequence prediction. By leveraging angle and dihedral relationships between backbone atoms, we ensure an E(3)-invariant representation of peptide structures. Our results demonstrate that our method achieves comparable performance to state-of-the-art models, highlighting its potential in pocket-aware peptide design. This work offers a new approach for precise drug discovery using receptor-specific peptide generation.

What problem does this paper attempt to address?

### What problem does this paper attempt to solve? This paper aims to solve **the challenges encountered in the discovery process of peptide inhibitors**, especially how to generate peptide structures and sequences for specific protein pockets. Specifically: 1. **Limitations of traditional peptide inhibitor discovery**: - **Labor - intensive and time - consuming**: The traditional peptide inhibitor discovery process is very cumbersome and time - consuming. - **Lack of specificity**: Most existing computer - aided peptide discovery methods mainly focus on generating peptide sequences or structures with general properties, ignoring the information of specific target pockets. 2. **Deficiencies of existing methods**: - **RFdiffusion model**: Although it can generate peptide structures around the target protein pocket residues, the subsequent inverse - folding models (such as ESM - IF or ProteinMPNN) lack information about pocket residues, which increases the risk that the generated peptides have high affinity for various protein structures rather than the specific target. - **Dataset limitations**: Existing inverse - folding models are usually trained on a wide range of protein complex datasets, which contain long amino acid sequences, limiting their specificity in the peptide inhibitor generation task. 3. **The proposed new method**: - **E(3)-invariant diffusion model**: The authors propose a new method to generate peptide structures and sequences through two sequential diffusion models. These two models are respectively used for end - to - end structure generation and sequence prediction, and ensure that the generated peptide structures are E(3)-invariant (i.e., invariant to rotation and translation in 3D space). - **Combining pocket information**: This method first generates peptide structures according to the shape of the pocket, and then combines the peptide structure and comprehensive pocket information when generating the sequence, thereby generating peptides that are more in line with specific target pockets. 4. **Objectives**: - **Improve the efficiency and accuracy of peptide inhibitor discovery**: By using deep learning and diffusion models, this method can generate peptide inhibitors for specific protein pockets more intelligently and efficiently. - **Provide practical tools**: Provide biologists with a more practical tool to help them find effective peptide inhibitors when studying biological processes and solving problems in agriculture, medicine and other fields. ### Summary The main objective of this paper is to achieve more accurate and efficient peptide inhibitor generation by proposing a new E(3)-invariant diffusion model and combining the information of specific protein pockets. This method not only improves the accuracy and specificity of peptide generation, but also provides a new approach for drug discovery.

E(3)-invariant diffusion model for pocket-aware peptide generation

A Unified Conditional Diffusion Framework for Dual Protein Targets Based Bioactive Molecule Generation

PIDiff: Physics informed diffusion model for protein pocket-specific 3D molecular generation

Interaction-based Retrieval-augmented Diffusion Models for Protein-specific 3D Molecule Generation

Functional-Group-Based Diffusion for Pocket-Specific Molecule Generation and Elaboration

Protein-Ligand Interaction Prior for Binding-aware 3D Molecule Diffusion Models

Pocket2Mol: Efficient Molecular Sampling Based on 3D Protein Pockets.

Pocket-specific 3D Molecule Generation by Fragment-based Autoregressive Diffusion Models

DiffBP: Generative Diffusion of 3D Molecules for Target Protein Binding

Complex-based Ligand-Binding Proteins Redesign by Equivariant Diffusion-based Generative Models

A Hybrid Diffusion Model for Stable, Affinity-Driven, Receptor-Aware Peptide Generation

Protein A-like Peptide Design Based on Diffusion and ESM2 Models

Full-Atom Peptide Design with Geometric Latent Diffusion

Structure-based Drug Design with Equivariant Diffusion Models

3D Equivariant Diffusion for Target-Aware Molecule Generation and Affinity Prediction

A pocket-based 3D molecule generative model fueled by experimental electron density

In-Pocket 3D Graphs Enhance Ligand-Target Compatibility in Generative Small-Molecule Creation

Hunting for peptide binders of specific targets with data-centric generative language models

Generating Novel, Designable, and Diverse Protein Structures by Equivariantly Diffusing Oriented Residue Clouds

Generalized Protein Pocket Generation with Prior-Informed Flow Matching

Protein generation with evolutionary diffusion: sequence is all you need