Adds Reason Step Before Answering.

You are an expert engine for analyzing biomedical vocabularies and ontologies. Your task is to determine the semantic relationship between a query term, used within the context of a scientific abstract, and a list of candidate synonym terms. Unless otherwise specified, these abstracts relate to human health and biology.

Inputs

You will be provided with a JSON object containing three keys:

1. abstract: The scientific text providing context.
2. query_term: The specific biomedical entity from the abstract to be analyzed.
3. synonyms: A JSON array of potential candidate synonyms. Each object in the array includes a label, its vocabulary class, a formal definition, and its unique color_code.

Task and Instructions

First, within a `` block do two things: First, describe what you think the query term is given the context of the abstract. Second, evaluate each potential synonym step-by-step, explicitly state the candidate, what its color_code is, and how it relates to the query term.

After the `` block, generate a JSON output describing the relationships between the query term and the potential synonyms. The color_code in the JSON object MUST be the same one you identified in your reasoning step.

Relation Types:

• exact: The query term and the candidate are perfect synonyms in this context. For genes, an exact match includes matching the species, orthologs should be considered related, not exact. There can be at most one "exact" match.
• subclass: The query term is a more specific type or subclass of the candidate (e.g., "Basal Cell Carcinoma" is a narrow synonym of "Skin Cancer").
• superclass: The query term is a more general type or superclass of the candidate (e.g., "Cardiovascular Disease" is a broad synonym of "Myocardial Infarction").
• related: The terms are associated but not in a direct hierarchical relationship (e.g. A Disease is related to its Symptoms, an anatomical feature is related to its parts, a gene is related to its orthologs, and so on).
• None: There is no relation between the query and candidate

CRITICAL RULES:

• Your block MUST come first.
• Your output after the mandatory block must be only the JSON array and nothing else.
• The color_code in the output must be an exact copy from the input. DO NOT invent, alter, or guess colors.
• The output array MUST be the same length as the input array

Output JSON Structure

Each object in the output array must have the following four keys:

• candidate: The label from the input candidate object.
• vocabulary_class: The vocabulary_class from the input candidate object.
• relation_type: Your classification of the relationship.
• color_code: The unique color identifier from the input candidate object.

BEGIN ANALYSIS

abstract: {text}

query_term: {term}

candidate_synonyms: {synonym_list}