A2A: a platform for research in biomedical literature search

Area of Science:

• Biomedical Informatics
• Information Retrieval
• Computational Biology

Background:

• Effective literature searching is vital for biomedical research and evidence-based medicine.
• Current search engines like PubMed offer limited user control over query and document processing.
• Biomedical researchers and computer scientists need enhanced control for information retrieval tasks.

Purpose of the Study:

• Introduce a public online tool for searching biomedical literature with greater user control.
• Facilitate exploration of different search strategies for biomedical information retrieval.
• Provide a platform guided by NIST TREC evaluation tasks in genomics, clinical decision support, and precision medicine.

Main Methods:

• Developed the A2A search and benchmarking tool.
• Utilized public datasets for evaluations.
• Implemented query formulation strategies and evaluated them against human judgments.

Main Results:

• Presented benchmark results for common biomedical information retrieval strategies.
• Evaluated strategies like querying MeSH subject headings and title-only searches.
• Reported standard information retrieval metrics, including precision at a cutoff.

Conclusions:

• The A2A tool is publicly available for researchers to explore novel search strategies.
• The tool supports exploration of diverse topics, from genomics to precision medicine.
• Evaluations demonstrate the utility of various query formulation approaches.