Targeted AMP generation through controlled diffusion with efficient embeddings

2025-04-24

AI Summary

To address low experimental hit rates, poor controllability, and inefficient modeling in antimicrobial peptide (AMP) discovery, this paper proposes OmegAMP—a controllable diffusion-based generative framework. Methodologically, it introduces a novel low-dimensional learnable embedding–driven multi-objective controllable diffusion mechanism, enabling targeted generation guided by physicochemical properties, broad- or narrow-spectrum activity, and species specificity. Additionally, a low false-positive ensemble classifier is designed to jointly optimize constraint satisfaction and sequence distribution fidelity. Across the full AMP generation pipeline, OmegAMP achieves state-of-the-art performance: experimental hit rate improves significantly; species-targeting accuracy increases by 23.6%; and both diversity and fidelity of generated sequences are enhanced concurrently. This work establishes a new paradigm for efficient discovery of anti–drug-resistant infection therapeutics.

Technology Category

• Search and Optimization: Sampling/Simulation-based Search
• Natural Language Processing: Generation
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

Deep learning-based antimicrobial peptide (AMP) discovery faces critical challenges such as low experimental hit rates as well as the need for nuanced controllability and efficient modeling of peptide properties. To address these challenges, we introduce OmegAMP, a framework that leverages a diffusion-based generative model with efficient low-dimensional embeddings, precise controllability mechanisms, and novel classifiers with drastically reduced false positive rates for candidate filtering. OmegAMP enables the targeted generation of AMPs with specific physicochemical properties, activity profiles, and species-specific effectiveness. Moreover, it maximizes sample diversity while ensuring faithfulness to the underlying data distribution during generation. We demonstrate that OmegAMP achieves state-of-the-art performance across all stages of the AMP discovery pipeline, significantly advancing the potential of computational frameworks in combating antimicrobial resistance.

Problem

• Enables targeted AMP generation with specific properties
• Improves controllability and reduces false positive rates
• Maximizes diversity while maintaining data distribution fidelity

Innovation

• Diffusion-based generative model with embeddings
• Precise controllability for specific AMP properties
• Novel classifiers reducing false positive rates