This research introduces Chemical Language Model Linker (CLM-Linker), a novel approach that connects large language models with molecular representation models for improved drug discovery.

• Leverages existing pre-trained models instead of building from scratch
• Uses modular adapters to efficiently bridge text and molecular data
• Achieves strong performance in generating viable drug candidates
• Enables creation of membrane-permeable protein inhibitors from text descriptions

For the medical industry, CLM-Linker offers a paradigm shift from large-scale chemical screening to targeted molecule generation, potentially accelerating drug development processes while reducing costs.

Chemical Language Model Linker: blending text and molecules with modular adapters