ModuLM: Enabling Modular and Multimodal Molecular Relational Learning with Large Language Models

2025-06-01

AI Summary

Current molecular relation learning (MRL) suffers from three key challenges: fragmented model architectures, rigid input modality constraints, and inconsistent, non-uniform evaluation protocols. To address these, we propose MolLLM—the first modular large language model (LLM) framework supporting multimodal molecular inputs (2D graphs and 3D conformations) with dynamic architecture switching. MolLLM decouples the encoder (8 × 2D + 11 × 3D variants), interaction layer (7 designs), and LLM backbone (7 models), enabling over 50,000 zero-redundancy composable configurations. It unifies molecular representation learning and relational reasoning within a single, extensible architecture. Empirically, MolLLM significantly improves model reusability and evaluation consistency across diverse MRL tasks, enabling fair, cross-architecture benchmarking. This work establishes a standardized, scalable LLM-MRL infrastructure for AI for Science.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Multi-modal Vision
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

Molecular Relational Learning (MRL) aims to understand interactions between molecular pairs, playing a critical role in advancing biochemical research. With the recent development of large language models (LLMs), a growing number of studies have explored the integration of MRL with LLMs and achieved promising results. However, the increasing availability of diverse LLMs and molecular structure encoders has significantly expanded the model space, presenting major challenges for benchmarking. Currently, there is no LLM framework that supports both flexible molecular input formats and dynamic architectural switching. To address these challenges, reduce redundant coding, and ensure fair model comparison, we propose ModuLM, a framework designed to support flexible LLM-based model construction and diverse molecular representations. ModuLM provides a rich suite of modular components, including 8 types of 2D molecular graph encoders, 11 types of 3D molecular conformation encoders, 7 types of interaction layers, and 7 mainstream LLM backbones. Owing to its highly flexible model assembly mechanism, ModuLM enables the dynamic construction of over 50,000 distinct model configurations. In addition, we provide comprehensive results to demonstrate the effectiveness of ModuLM in supporting LLM-based MRL tasks.

Problem

• Lack of flexible LLM framework for molecular input formats
• No support for dynamic architectural switching in MRL
• Challenges in benchmarking diverse LLMs and encoders

Innovation

• Modular framework for flexible LLM-based model construction
• Supports diverse 2D and 3D molecular encoders
• Enables dynamic assembly of 50,000+ model configurations