FragmentGPT: A Unified GPT Model for Fragment Growing, Linking, and Merging in Molecular Design

2025-09-13

AI Summary

Fragment-based drug discovery (FBDD) faces challenges in designing linkers for fragment connectors and suffers from structural redundancy—such as redundant ring systems—that compromise molecular feasibility. Method: This paper introduces the first unified framework that jointly addresses fragment growth, linking, and merging. It innovatively proposes a chemistry-aware, energy-based bond cleavage pretraining strategy, integrated with a reward-ranking expert alignment algorithm and Pareto-optimal multi-objective optimization. Within a GPT architecture, it synergistically combines expert imitation learning, data augmentation, and supervised fine-tuning for conditional generation of linker molecules. Contribution/Results: The framework generates molecules with high bioactivity, synthetic accessibility, and structural diversity. Evaluated on real-world cancer target datasets, it achieves 98.7% chemical validity and improves diversity by 2.1× over state-of-the-art methods, demonstrating significant utility for downstream drug discovery.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Search and Optimization: Learning to Search
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• 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

Abstract

Fragment-Based Drug Discovery (FBDD) is a popular approach in early drug development, but designing effective linkers to combine disconnected molecular fragments into chemically and pharmacologically viable candidates remains challenging. Further complexity arises when fragments contain structural redundancies, like duplicate rings, which cannot be addressed by simply adding or removing atoms or bonds. To address these challenges in a unified framework, we introduce FragmentGPT, which integrates two core components: (1) a novel chemically-aware, energy-based bond cleavage pre-training strategy that equips the GPT-based model with fragment growing, linking, and merging capabilities, and (2) a novel Reward Ranked Alignment with Expert Exploration (RAE) algorithm that combines expert imitation learning for diversity enhancement, data selection and augmentation for Pareto and composite score optimality, and Supervised Fine-Tuning (SFT) to align the learner policy with multi-objective goals. Conditioned on fragment pairs, FragmentGPT generates linkers that connect diverse molecular subunits while simultaneously optimizing for multiple pharmaceutical goals. It also learns to resolve structural redundancies-such as duplicated fragments-through intelligent merging, enabling the synthesis of optimized molecules. FragmentGPT facilitates controlled, goal-driven molecular assembly. Experiments and ablation studies on real-world cancer datasets demonstrate its ability to generate chemically valid, high-quality molecules tailored for downstream drug discovery tasks.

Problem

• Designing effective linkers for fragment-based drug discovery
• Resolving structural redundancies in molecular fragments
• Optimizing molecules for multiple pharmaceutical goals

Innovation

• Chemically-aware bond cleavage pre-training strategy
• Reward Ranked Alignment with Expert Exploration algorithm
• Generates linkers connecting fragments optimizing pharmaceutical goals