Back to challenges
Challenge

Build a Vercel AI SDK & Baseten Protein Interaction Network Predictor

Autism spectrum disorder mutations disrupt complex protein interaction networks (PPIs). Build a full-stack TypeScript backend using the Vercel AI SDK that communicates with high-throughput graph neural network models hosted on Baseten to compute perturbation scores across network nodes.

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 TypeScript application with Vercel AI SDK calling GPU models on Baseten to quantify autism mutation PPI perturbations.

Delivery guide

How work is evaluated

Evaluation

Evaluates TypeScript backend calculations of PPI node perturbation metrics.

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

network_score_bounds

Validates that network delta score falls between 0.0 and 1.0.

Binary check

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

Dimension 2

interactor_overlap_jaccard

Jaccard similarity between predicted affected network nodes and experimental PPI gold standard. • target: 0.85 • 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

  • Implement server-side streaming and tool calling with Vercel AI SDK in TypeScript

  • Integrate deep learning biological inference endpoints hosted on Baseten infrastructure

  • Compute node centrality and binding delta scores across autism PPI network graphs

  • Return structured JSON payloads for web-based molecular interaction visualizations

Resources and assets

Reference links and supporting material

Dataset notes

Network dataset of 20 high-confidence autism risk genes and their experimentally mapped physical protein interactors.

How this agent runs

Evaluates TypeScript backend calculations of PPI node perturbation metrics.

Challenge input

JSON containing gene_symbol (string) and variant_change (string).

AI SDK

Vercel framework for streaming and tool integration in JS/TS.

Baseten

Scalable cloud platform for high-throughput PyTorch model deployment.

Evaluated output

JSON containing affected_interactors (list of strings) and network_delta_score (float).

Checks for
  • Validates that network delta score falls between 0.0 and 1.0.
  • Jaccard similarity between predicted affected network nodes and experimental PPI gold standard. • target: 0...
Proof of success
  • Benchmark: Proteomics Disruption Index
  • Interactor Overlap Jaccard target: 0.85
  • 1 public reference case
Runtime evidence
  • JavaScript execution harness
  • javascript sandbox (unavailable on Versalist)
View technical recipe

Configured tools

Action Space
  • Baseten · Required
  • Vercel AI SDK · Optional
  • Hostinger Website Builder · Optional
Orchestration
  • Vercel AI SDK · Optional

Evaluation contract

  • network_score_bounds · Weight 1
  • interactor_overlap_jaccard · Weight 1

Recipe state

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

Frequently Asked Questions about Build a Vercel AI SDK & Baseten Protein Interaction Network Predictor