Pepti-Agent: An AI Agent for Peptide Design and Optimization

2026-06-13

AI Summary

This work addresses the challenge of multi-objective peptide design, where key properties such as solubility, hemolytic activity, and resistance to non-specific adsorption are often mutually conflicting and difficult to optimize simultaneously. To tackle this, the authors propose a closed-loop, interpretable AI agent framework that, for the first time, employs the Model Context Protocol to decouple generation, prediction, and editing modules. The framework dynamically orchestrates a large language model to coordinate PeptideGPT for sequence generation, ProtBERT for multi-property classification, and a single-residue mutation operator. It enables full traceability of decision-making processes and strategy reproducibility, thereby achieving transparent and reproducible multi-objective peptide optimization and providing a reliable basis for prioritizing experimental peptide candidates.

Abstract

Therapeutic peptides occupy a valuable design space between small molecules and biologics, but their development requires satisfying several competing constraints at once: solubility, hemolytic activity, and nonspecific surface fouling are governed by overlapping sequence features, so improving one property often degrades another. Computational design addresses this by pairing generative models with sequence-based property predictors, iteratively proposing and refining candidates. However, these components are typically wired together as monolithic scripts that are difficult to inspect, extend, or reuse, and they often refine sequences by natural-language reasoning rather than by tracking the evolving multi-property state of each candidate. We present Pepti-Agent, a closed-loop, peptide-specific framework that exposes generation, property prediction, and single-residue mutation as independently inspectable Model Context Protocol (MCP) tools. A large language model controller invokes these tools and consults live predictor output between calls, so refinement is guided by each sequence's current property profile rather than by language reasoning alone. Task-specific PeptideGPT models generate candidates, ProtBERT-based classifiers score solubility, hemolysis, and non-fouling, and two interchangeable mutation operators propose sequence edits. By recording a per-step trace of controller decisions, predictor outputs, and accepted mutations, Pepti-Agent offers a reproducible substrate for benchmarking multi-objective design strategies and for prioritizing candidates for experimental validation.

Problem

• peptide design
• multi-objective optimization
• competing constraints
• property prediction
• sequence optimization

Innovation

• Pepti-Agent
• Model Context Protocol
• multi-objective peptide design
• large language model controller
• property-guided optimization