Genolator enables protein function interpretation using a multimodal large language model fusing genomic and structural interpretation with natural language interaction.

Genolator, a multimodal large language model, achieves over 365,000 question-answer pairs to accurately interpret protein functions from genomic and structural data.

• Why it matters: Decoding the genome's complex language is crucial for understanding disease mechanisms and developing targeted treatments, but current models lack integration of genomic and structural information for natural language interaction.
• What they did: The team developed Genolator by fine-tuning a multimodal LLM on diverse embeddings from DNA, amino acid sequences, and protein structures, enabling it to answer questions about protein localization, function, and processes.
• The result: Genolator outperforms existing models like GPT 4.1, providing high-accuracy, biologically plausible insights into protein functions, thus advancing accessible genomic interpretation for biological and clinical research.