TourSynbio-Search: LLM Agent for Unified Protein Engineering Search

The rapid expansion of protein-related databases and scientific literature has created a significant challenge for researchers seeking efficient information retrieval in protein engineering. Traditional methods often require specialized bioinformatics expertise, creating a barrier to accessing and analyzing complex biological data. To address this, the TourSynbio-Search framework has been developed. Powered by the TourSynbio-7B protein multimodal large language model, this agent system features a dual-module architecture, PaperSearch and ProteinSearch, allowing comprehensive exploration of both scientific literature (e.g., ArXiv, BioRxiv) and protein data (e.g., UniProt, PDB). Its core innovation lies in interpreting natural language queries, optimizing search parameters, and executing operations across major platforms, thereby streamlining biological information retrieval and enhancing research productivity for biologists without extensive computational backgrounds.