Calchas

Calchas enables semantic-based discovery of biomedical resources by translating natural-language clinical queries into ontology-based searches to identify and rank relevant computational tools and resources.

Key Features:

• Semantic Annotation: Uses domain-specific biomedical ontologies to semantically annotate resource descriptions and align them with established medical terminologies and concepts.
• Natural Language Processing (NLP): Transforms free-text clinical queries into targeted ontology-term queries using information extraction techniques and integrated ontologies.
• Automated Resource Discovery: Maps query descriptions onto domain ontologies and ranks available resources by similarity matches to prioritize relevance.
• Precision and Recall Evaluation: Assesses retrieval performance using precision and recall metrics, reporting high precision and low recall in evaluations.

Scientific Applications:

• Clinical research tool identification: Identifies computational pipelines or individual tools that address specific clinical decision tasks based on ontology-aligned query matching.
• Natural-language-to-ontology bridging: Enables clinicians and researchers to express clinical research questions in natural language and retrieve ontology-grounded biomedical resources.
• Benchmarking with ClinicalTrials.gov: Supports evaluation and benchmarking of resource discovery using clinical questions derived from ClinicalTrials.gov and clinician recommendations.

Methodology:

Integration of biomedical ontologies, NLP-based query translation, text mining to map resource descriptions onto domain ontologies, and evaluation through experiments using ClinicalTrials.gov-derived clinical questions with clinician recommendations and domain-expert manual identification of suitable tools assessed by precision and recall.