Skip to main content
Back to challenges
Challenge

Drug Toxicity Cell-Free DNA Monitoring System using CrewAI and Google Vertex AI

Monitoring early drug-induced organ toxicity requires continuous evaluation of cell-free DNA methylation markers in blood samples. Build a CrewAI multi-agent workflow integrated with Google Vertex AI to assess cfDNA tissue-of-origin signals and trigger drug safety alerts within 15 seconds per batch.

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.

Build an auditable drug toxicity alerting agent pipeline with CrewAI and Google Vertex AI model endpoints.

Delivery guide

How work is evaluated

Evaluation

Evaluates latency and hepatotoxicity detection accuracy from cfDNA methylation profiles.

Datasets

Challenge datasets

Public datasets, plus the private datasets you have access to.

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

alert_logic_test

Ensures toxicity_flag triggers correctly when liver fraction > 0.10.

Binary check

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

Dimension 2

latency_seconds

Batch processing response time. • target: 15 • range: 0-30

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

  • Orchestrate role-playing CrewAI agents for genomic signal processing.

  • Integrate Google Vertex AI models for fast sequence annotation.

  • Deconvolve cfDNA methylation markers to determine tissue-of-origin.

  • Automate drug safety reporting when hepatotoxicity thresholds are exceeded.

Resources and assets

Reference links and supporting material

Dataset notes

Simulated cfDNA tissue-of-origin methylation fractions across blood samples collected during Phase I clinical drug trials.

Challenge setup

Evaluates latency and hepatotoxicity detection accuracy from cfDNA methylation profiles.

Challenge input

JSON containing tissue methylation fractions.

CrewAI

Required multi-agent framework for collaborative task execution.

Google Vertex AI

Scalable enterprise LLM runtime for bio-entity analysis.

Evaluated output

JSON containing toxicity_flag and tissue_origin_scores.

Evaluation checks
  • Ensures toxicity_flag triggers correctly when liver fraction > 0.10.
  • Batch processing response time. • target: 15 • range: 0-30
Configured evaluation evidence
  • Benchmark: Pharmacovigilance cfDNA Alert Benchmark
  • Latency Seconds target: 15
  • 1 public reference case
Configured runtime
  • 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

Action Space
  • Google · Required
  • Exa · Optional
  • Google Vertex AI · Optional

Evaluation contract

  • alert_logic_test · Weight 1
  • latency_seconds · Weight 1

Metrics

  • latency_seconds · Target 15

Benchmarks

  • Pharmacovigilance cfDNA Alert Benchmark
  • biotech

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON containing tissue methylation fractions.

    Output: JSON containing toxicity_flag and tissue_origin_scores.

    Execution harness: Python

Evaluation tools

  • CrewAI · Agent Framework
  • Google Vertex AI · Inference Runtime

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Drug Toxicity Cell-Free DNA Monitoring System using CrewAI and Google Vertex AI