EMMAA Architecture and Approach

The Ecosystem of Machine-maintained Models with Automated Analysis is available at http://github.com/indralab/emmaa, with the EMMAA Model Dashboard at http://emmaa.indra.bio.

The main idea behind EMMAA is to create a set of computational models that are kept up-to-date using automated machine reading, knowledge-assembly, and model generation, integrating new discoveries immediately as they become available.

As a key component of the approach, models are automatically tested against experimental observations (also automatically gathered and associated with models). Models are also available for automated analysis in which relevant queries that fall within the scope of each model can be automatically mapped to structural and dynamical analysis procedures on the model. Currently, the Dashboard supports running and registering queries with respect to one or more existing models. In the near future, EMMAA will automatically recognize and report changes to each model that result in meaningful changes to analysis results.

• Model Assembly and Updates
  • Cancer types of interest
  • Model availability
  • Defining model scope
  • Deriving relevant terms for a given type of cancer
    • Finding disease genes
    • Finding relevant entities in a knowledge network
    • Assembling a prior network
  • Updating the network
  • Machine-reading
  • Automated incremental assembly
• Meta-Model
  • Initial specification of annotation guidelines
  • EMMAA currently supports “does X…” queries for PySB models
  • Annotations required for “what if” queries
  • Annotations required for open-ended “relevance” queries
• Model Testing and Analysis
  • Model test cycle deployed on AWS
  • Test conditions generated automatically
  • General EMMAA model testing framework
    • Test conditions mapped to models automatically
    • Testing models using static analysis
    • Human-readable model test reports
  • Model queries from users
  • Pre-registered queries and notifications
• Model Analysis Query Language
  • Structural properties with constraints
    • Specifying topology
    • Entity constraints
    • Relationship constraints
    • Examples
    • Ideas for extension
  • Path properties with constraints
    • Specifying the overall path
    • Entity constraints
    • Relationship constraints
    • Examples
  • Simple intervention properties
    • Specifying an intervention
    • Specifying the reference
    • Specifying the intervetion
    • Examples
  • Comparative intervention properties
    • Examples