Pipeline to explore information on genome editing using large language models and genome editing meta-database

Area of Science:

• Life Sciences
• Bioinformatics
• Genomics

Background:

• Genome editing (GE) is a powerful tool in life sciences, but gene editability varies by species, sequence, and tools.
• Identifying prior GE applications is crucial for designing new GE research.
• Existing databases like GEM lack detailed gene-specific GE involvement information.

Purpose of the Study:

• To develop a systematic method for extracting essential genome editing information from scientific literature.
• To enhance the utility of existing databases by providing gene-specific GE involvement data.
• To propose a metric-based system for prioritizing genes for future genome editing research.

Main Methods:

• Utilized large language models to systematically extract genome editing information from the Genome Editing Meta-database (GEM) and related articles.
• Developed a method to analyze and process extracted data for efficient information retrieval.
• Converted extracted GE data into quantifiable metrics for gene prioritization.

Main Results:

• Successfully developed a systematic approach to extract detailed gene-specific genome editing information.
• The method enables more efficient and comprehensive investigation of GE data than current resources alone.
• Generated novel GE-related scores to aid in prioritizing genes for future research.

Conclusions:

• The developed large language model-based method significantly improves the extraction and analysis of genome editing information.
• This approach facilitates efficient selection of target genes and supports the design of advanced genome editing studies.
• The generated metrics offer a valuable tool for prioritizing genes in future GE research endeavors.