Automated extraction and optimization of protein purification protocols using multi-agent large language models

Area of Science:

• Biotechnology
• Computational Biology
• Biochemistry

Background:

• Protein purification is critical for biomedical research but often fails.
• Manual optimization of purification protocols is time-intensive and complex.
• Existing methods lack efficiency in handling proteins with low stability or solubility.

Purpose of the Study:

• To automate the creation and optimization of protein purification protocols using a multi-agent Large Language Model (LLM) system.
• To improve the success rate of producing high-concentration, high-purity protein samples.
• To streamline the manual processes involved in sequence similarity searches, literature reviews, and protocol comparison.

Main Methods:

• A multi-agent LLM system was developed to operate within a constrained workflow.
• The system identifies analogous proteins and extracts purification methodologies from scientific literature.
• It cross-references successful and failed protocols to generate optimization recommendations.

Main Results:

• The system demonstrated high accuracy in extracting purification protocols from literature.
• It generated scientifically sound and expert-validated optimization recommendations.
• Complex analysis time was reduced from hours to minutes for target proteins.

Conclusions:

• LLM agents can effectively automate and streamline complex wet-lab workflows like protein purification.
• The developed system enhances methodological transparency and reproducibility in scientific research.
• Programmatic open access to scientific literature, particularly primary citations, is crucial for advancing LLM-based scientific workflows.