Skip to main content
Back to challenges
Challenge

Mastra AI Protein Interaction Network Agent with fal.ai

Autism spectrum disorder mutations alter complex protein-protein interaction (PPI) networks and structural topologies. Build a TypeScript workflow using Mastra AI integrated with fal.ai inference models to evaluate structural impact metrics for missense variants in PPI nodes. Achieve structural prediction latency under two seconds per mutation.

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 agent workflow using Mastra AI and fal.ai fast inference endpoints to evaluate ASD missense mutation effects on protein interaction interfaces.

Delivery guide

How work is evaluated

Evaluation

Evaluates response latency and topological disruption scoring precision of the Mastra workflow.

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

latency_under_2s

Checks if task processing completes in under 2000 milliseconds.

Binary check

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

Dimension 2

ddg_correlation

Pearson correlation coefficient with experimental ddG values. • target: 0.82 • range: -1-1

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 a Mastra AI Agent workflow in Node.js/TypeScript

  • Connect fal.ai protein structure inference models via REST API

  • Calculate interface delta binding energy (ddG) and node centrality metrics

  • Generate interactive PPI network alteration graphs

Resources and assets

Reference links and supporting material

Dataset notes

Structural dataset of ASD-associated mutations annotated with experimental binding affinity changes and network interactome annotations.

Challenge setup

Evaluates response latency and topological disruption scoring precision of the Mastra workflow.

Challenge input

JSON containing protein_id, wildtype, mutation_pos, mutant

Mastra AI

Required framework for building TypeScript workflows.

fal.ai

Provides low-latency structural prediction APIs.

Evaluated output

JSON with predicted_ddg, network_disruption_score, processing_time_ms

Evaluation checks
  • Checks if task processing completes in under 2000 milliseconds.
  • Pearson correlation coefficient with experimental ddG values. • target: 0.82 • range: -1-1
Configured evaluation evidence
  • Benchmark: PPIPredict-Bench
  • Ddg Correlation target: 0.82
  • 1 public reference case
Configured runtime
  • JavaScript execution harness
  • javascript 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
  • fal.ai · Required
  • Mastra AI · Optional
  • Builder.ai · Optional
Orchestration
  • Mastra AI · Optional

Evaluation contract

  • latency_under_2s · Weight 1
  • ddg_correlation · Weight 1

Metrics

  • ddg_correlation · Target 0.82

Benchmarks

  • PPIPredict-Bench
  • biotech

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON containing protein_id, wildtype, mutation_pos, mutant

    Output: JSON with predicted_ddg, network_disruption_score, processing_time_ms

    Execution harness: JavaScript

Evaluation tools

  • Mastra AI · TypeScript Agent framework
  • fal.ai · Model inference host

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Mastra AI Protein Interaction Network Agent with fal.ai