Adjutant: an R-based tool to support topic discovery for systematic and literature reviews

Area of Science:

• Bioinformatics
• Computational Biology
• Data Mining

Background:

• Literature reviews are crucial for synthesizing research findings.
• Efficiently mining large biomedical literature databases like PubMed is challenging.
• Identifying key themes and topics within extensive datasets requires robust analytical tools.

Purpose of the Study:

• To introduce Adjutant, an open-source R-based application designed to streamline the process of mining PubMed for literature reviews.
• To provide researchers with a tool for unsupervised clustering analysis to discover data-driven topic clusters within downloaded articles.
• To enable efficient data sampling for creating manageable datasets for further in-depth analysis.

Main Methods:

• Adjutant utilizes an R-based framework with a Shiny interface for interactive use.
• It accepts PubMed-compatible search queries to download relevant articles.
• Implements unsupervised clustering algorithms for topic identification and offers various document sampling strategies.

Main Results:

• Adjutant facilitates the identification of data-driven topic clusters from PubMed literature.
• The application allows for adjustable trade-offs between analysis speed and accuracy.
• Complex analyses of thousands of documents can be completed within minutes.
• Generated analytic datasets are saved for downstream analyses.

Conclusions:

• Adjutant offers an efficient and interactive solution for researchers conducting literature reviews using PubMed.
• The tool supports data-driven topic discovery and dataset management for biomedical research.
• Its flexibility in speed-accuracy trade-offs and data export capabilities enhance research workflows.