Back to evaluations
Draft evaluation
Build a Variant Functional Effect Predictor with OpenAI Agents SDK — evaluation
Evaluates variant triage accuracy, ACMG code assignment precision, and JSON output formatting.
Evaluation type
task based
Challenge
Build a Variant Functional Effect Predictor with OpenAI Agents SDK
Difficulty
Intermediate
Rigor
Not declared
The author has not specified a rigor level.
Evaluation overview
How the linked challenge is judged: tasks, benchmarks, and criteria count.
Tasks
1
Benchmarks
0
Criteria
0
Task templates
Inputs and expected outputs.
Task 1
variant_classification_task
Evaluates the agent's ability to classify missense variants into ACMG categories.
Input format
JSON containing genomic variant coordinates and reference/alternate alleles.
Output format
JSON containing ACMG classification, pathogenicity score, and evidence codes.