MolTextNet: A Two-Million Molecule-Text Dataset for Multimodal Molecular Learning
This addresses a data bottleneck for researchers in drug discovery and molecular science who need generalizable multimodal models.
The authors tackled the problem of limited scale and informativeness in existing molecule-text datasets by creating MolTextNet, a dataset of 2.5 million high-quality molecule-text pairs, which enabled pretraining multimodal models that showed improved performance on downstream tasks.
Small molecules are essential to drug discovery, and graph-language models hold promise for learning molecular properties and functions from text. However, existing molecule-text datasets are limited in scale and informativeness, restricting the training of generalizable multimodal models. We present MolTextNet, a dataset of 2.5 million high-quality molecule-text pairs designed to overcome these limitations. To construct it, we propose a synthetic text generation pipeline that integrates structural features, computed properties, bioactivity data, and synthetic complexity. Using GPT-4o-mini, we create structured descriptions for 2.5 million molecules from ChEMBL35, with text over 10 times longer than prior datasets. MolTextNet supports diverse downstream tasks, including property prediction and structure retrieval. Pretraining CLIP-style models with Graph Neural Networks and ModernBERT on MolTextNet yields improved performance, highlighting its potential for advancing foundational multimodal modeling in molecular science. Our dataset is available at https://huggingface.co/datasets/liuganghuggingface/moltextnet.