PheKnow-Cloud: A Tool for Evaluating High-Throughput Phenotype Candidates using Online Medical Literature

Area of Science:

• Biomedical Informatics
• Computational Biology
• Health Informatics

Background:

• The increasing use of Electronic Healthcare Records (EHRs) necessitates automated methods for data extraction and characterization.
• Current methods for extracting candidate phenotypes from EHRs often rely on unsupervised or semi-supervised approaches, requiring clinical validation.
• Assessing the clinical relevance of automatically generated phenotypes is crucial for their reliable application.

Purpose of the Study:

• To introduce PheKnow-Cloud, a novel framework designed to support the evaluation of candidate phenotypes derived from EHR data.
• To facilitate the validation of automatically extracted phenotypes by providing robust evidence sets.
• To assist researchers and clinicians in understanding and refining phenotype extraction processes.

Main Methods:

• PheKnow-Cloud employs co-occurrence analysis on the PubMed database, a vast repository of biomedical literature.
• The framework processes user-supplied candidate phenotypes to identify supporting evidence within PubMed.
• Results are presented interactively, allowing for user-driven exploration and analysis of phenotype evidence.

Main Results:

• PheKnow-Cloud successfully builds evidence sets for candidate phenotypes by leveraging large-scale literature data.
• The interactive presentation of results enables efficient examination of the clinical context and support for phenotypes.
• The framework provides a mechanism to assess the validity and relevance of phenotypes generated from EHRs.

Conclusions:

• PheKnow-Cloud offers a valuable tool for the clinical validation of automatically extracted phenotypes from EHRs.
• By utilizing PubMed co-occurrence analysis, the framework enhances the reliability and interpretability of phenotype discovery.
• This approach aids in bridging the gap between automated data extraction and clinical utility in precision medicine.