Abstract:Denoising Diffusion Probabilistic Models (DDPMs) have emerged as a potent class of generative models, demonstrating exemplary performance across diverse AI domains such as computer vision and natural language processing. In the realm of protein design, while there have been advances in structure-based, graph-based, and discrete sequence-based diffusion, the exploration of continuous latent space diffusion within protein language models (pLMs) remains nascent. In this work, we introduce AMP-Diffusion, a latent space diffusion model tailored for antimicrobial peptide (AMP) design, harnessing the capabilities of the state-of-the-art pLM, ESM-2, to generate functional AMPs for downstream experimental application. Our evaluations reveal that peptides generated by AMP-Diffusion align closely in both pseudo-perplexity and amino acid diversity when benchmarked against experimentally-validated AMPs, and further exhibit relevant physicochemical properties similar to these naturally-occurring sequences. Overall, these findings underscore the biological plausibility of our generated sequences and pave the way for their empirical validation. In total, our framework motivates future exploration of pLM-based diffusion models for peptide and protein design.

What problem does this paper attempt to address?

The paper mainly discusses how to use deep learning methods to generate antimicrobial peptides (AMPs) with antibacterial functionality. Currently, although there has been progress in diffusion models based on structure, graph, and sequence, the study of continuous latent space diffusion within protein language models (pLMs) is still in its early stages. The paper proposes AMP-Diffusion, a latent space diffusion model specifically designed for AMPs, which combines the capabilities of the state-of-the-art pLM ESM-2 to generate functional AMPs from scratch for experimental validation. The AMP-Diffusion model introduces Gaussian noise into the latent representation of protein sequences, and then reconstructs the sequences by removing this noise through a reverse process. After training, the generated AMPs perform well in terms of pseudoperplexity, amino acid diversity, and similarity to experimentally validated AMPs. Furthermore, these AMPs also exhibit similar physicochemical properties to natural AMPs, such as charge, hydrophobicity, and isoelectric point, indicating the biological rationality of the generated sequences. The paper evaluates the performance of AMP-Diffusion compared to other models such as HydrAMP, PepCV AE, and AMPGAN, and demonstrates superiority across multiple metrics. Through the external classifier HydrAMP, the peptide segments generated by AMP-Diffusion also show high potential in terms of the probability of being classified as antimicrobial peptides and the likelihood of activity against Escherichia coli. In conclusion, this work fills the gap between pLMs and diffusion model integration, providing a new tool for AMP design and opening up new directions for future protein design research, including experimental validation of the activity of generated peptide segments, customization of peptide segments with specific properties, and extension of this method to a broader range of protein engineering applications.

AMP-Diffusion: Integrating Latent Diffusion with Protein Language Models for Antimicrobial Peptide Generation

ProT-Diff: A Modularized and Efficient Approach to De Novo Generation of Antimicrobial Peptide Sequences through Integration of Protein Language Model and Diffusion Model

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

TG-CDDPM: text-guided antimicrobial peptides generation based on conditional denoising diffusion probabilistic model

Diffusion on language model embeddings for protein sequence generation

PRO-LDM: Protein Sequence Generation with a Conditional Latent Diffusion Model

Diffusion Language Models Are Versatile Protein Learners

ProT‐Diff: A Modularized and Efficient Strategy for De Novo Generation of Antimicrobial Peptide Sequences by Integrating Protein Language and Diffusion Models

MeMDLM: De Novo Membrane Protein Design with Masked Discrete Diffusion Protein Language Models

Diff-AMP: Tailored Designed Antimicrobial Peptide Framework with All-in-one Generation, Identification, Prediction and Optimization.

Accelerating Antimicrobial Peptide Discovery with Latent Structure

DPLM-2: A Multimodal Diffusion Protein Language Model

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

Application of a deep generative model produces novel and diverse functional peptides against microbial resistance

A Latent Diffusion Model for Protein Structure Generation

Protein Language Models and Machine Learning Facilitate the Identification of Antimicrobial Peptides

Full-Atom Peptide Design with Geometric Latent Diffusion

Protein Language Model Supervised Precise and Efficient Protein Backbone Design Method

Exploring Latent Space for Generating Peptide Analogs Using Protein Language Models

Deep Learning for Novel Antimicrobial Peptide Design

MOFormer: Multi-Objective Antimicrobial Peptide Generation Through Conditional Transformer with Multi-Fusion Descriptors