Build a Newborn Genetic Risk Triage State Machine with LangChain & LangGraph
Newborn genomic screening can detect early childhood cancer risks, but requires deterministic state management during clinical follow-up triage. Build a stateful diagnostic workflow engine using LangChain and LangGraph to transition variant findings into clinical care pathways with 98% classification fidelity.
What you are building
The core problem, expected build, and operating context for this challenge.
Construct a stateful diagnostic triage pipeline using LangChain and LangGraph for newborn genomic screening.
How work is evaluated
Evaluates state transition accuracy and pathogenicity classification on newborn genetic screening profiles.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
pathogenic_priority_check
Ensures pathogenic variants trigger URGENT priority.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
triage_accuracy
Accuracy of clinical care pathway assignments. • target: 0.98 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Design cyclic state graph architectures using LangGraph `StateGraph`.
Incorporate ACMG/AMP variant classification logic inside LangChain graph nodes.
Route pediatric cancer risk variants into automated confirmatory workflow branches.
Reference links and supporting material
Dataset of 50 synthetic newborn genetic screening variant records with expert-annotated clinical triage labels.
Challenge setup
Evaluates state transition accuracy and pathogenicity classification on newborn genetic screening profiles.
Challenge input
JSON containing gene_symbol, hgvs_c, and clinical_significance
LangChain
Popular AI framework providing chain components and tool definitions.
LangGraph
State machine graph orchestration for complex cyclic logic.
Evaluated output
JSON object with classification, priority_level, and next_step
- Ensures pathogenic variants trigger URGENT priority.
- Accuracy of clinical care pathway assignments. • target: 0.98 • range: 0-1
- Benchmark: GenoTriage-Bench
- Triage Accuracy target: 0.98
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
- Langchain · Required
- LangChain · Optional
- Continue · Optional
- Langchain · Required
- LangChain · Optional
Evaluation contract
- pathogenic_priority_check · Weight 1
- triage_accuracy · Weight 1
Metrics
- triage_accuracy · Target 0.98
Benchmarks
- GenoTriage-Bench
- biotech
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON containing gene_symbol, hgvs_c, and clinical_significance
Output: JSON object with classification, priority_level, and next_step
Execution harness: Python
Evaluation tools
- LangChain · agent_framework
- LangGraph · graph_orchestrator
Recipe freeze marker
No recipe freeze time is recorded.