Back to challenges
Challenge

Build a LlamaIndex Pediatric Genomics Decision Agent on Amazon Bedrock

Pediatric rapid Whole-Genome Sequencing (rWGS) produces vast variant calls that require rapid mapping against Human Phenotype Ontology (HPO) terms. Construct a LlamaIndex decision support agent deployed on Amazon Bedrock that indexes variant annotations and provides phenotype-driven diagnostic candidates.

Data ScienceHosted by Vera
Sign in free to start
Challenge brief

What you are building

The core problem, expected build, and operating context for this challenge.

Develop a LlamaIndex RAG and decision agent on Amazon Bedrock to match pediatric rWGS genomic variants with HPO clinical terms.

Delivery guide

How work is evaluated

Evaluation

Evaluates candidate disease ranking accuracy given patient HPO phenotype terms and variant call inputs.

Datasets

Shared data for this challenge

Review public datasets and any private uploads tied to your build.

Evaluation rubric

How submissions are scored

These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.

Dimensions
2 scoring checks
Binary
2 pass or fail dimensions
Ordinal
0 scaled dimensions
Dimension 1

top_candidate_match

Verifies correct disease is ranked in top 3 positions.

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Dimension 2

mean_reciprocal_rank

MRR score of true diagnosis across rare disease benchmark test cases. • target: 0.88 • range: 0-1

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Learning goals

What you should walk away with

  • Build a LlamaIndex vector and graph index over HPO and genomic transcript databases

  • Integrate Amazon Bedrock Claude foundation models for high-throughput clinical reasoning

  • Design query engines capable of phenotype-to-genotype candidate disease ranking

  • Implement audit logging for reproducible clinical evidence chains

Resources and assets

Reference links and supporting material

Dataset notes

Cohort of 40 pediatric clinical cases containing HPO phenotype codes and associated pathogenic variant calls with verified diagnoses.

How this agent runs

Evaluates candidate disease ranking accuracy given patient HPO phenotype terms and variant call inputs.

Challenge input

JSON containing hpo_terms (list of strings) and detected_variants (list of strings).

LlamaIndex

Structures indices and retrieval tools.

Amazon Bedrock

Provides enterprise-grade LLM inference.

Evaluated output

JSON with ranked_diseases (list of objects with disease_id and score).

Checks for
  • Verifies correct disease is ranked in top 3 positions.
  • MRR score of true diagnosis across rare disease benchmark test cases. • target: 0.88 • range: 0-1
Proof of success
  • Benchmark: HPO Rare Disease Retrieval Benchmark
  • Mean Reciprocal Rank target: 0.88
  • 1 public reference case
Runtime evidence
  • Python execution harness
  • Python sandbox (unavailable on Versalist)
View technical recipe

Configured tools

Action Space
  • Amazon Bedrock · Required
  • Llama Index · Optional
  • MutableAI · Optional

Evaluation contract

  • top_candidate_match · Weight 1
  • mean_reciprocal_rank · Weight 1

Recipe state

This is a preview. The configuration can change before the evaluation recipe is locked.

Frequently Asked Questions about Build a LlamaIndex Pediatric Genomics Decision Agent on Amazon Bedrock