MPS

MPS enables creation of domain-specific languages using Language Workbench (LW) technology to model Dataset, Endpoint, Feature Selection Method, and Classifier abstractions and to support training and validation of predictive biomarker models from high-throughput datasets.

Key Features:

• Language Workbench (LW) support: Leverages Language Workbench (LW) technology to implement domain-specific languages for biological analyses.
• High-level concept modeling: Models abstractions including Dataset, Endpoint, Feature Selection Method, and Classifier.
• Model training and validation: Supports training predictive models and producing validation statistics.
• Language composition: Supports language composition to enable consistency, portability, and extensibility of domain-specific languages.
• BDVal integration: Integrates the BDVal plugin to configure BDVal projects for biomarker development.
• High-throughput dataset focus: Targets development of predictive models from high-throughput datasets.

Scientific Applications:

• Biomarker development and validation: Configuring, training, and validating predictive biomarker models from high-throughput datasets.
• Predictive model development: Developing classifiers and feature-selection workflows for predictive modeling using modeled Dataset and Endpoint abstractions.
• BDVal project configuration: Setting up BDVal projects specifically for biomarker development and validation.

Methodology:

Leverages Language Workbench (LW) technology to define domain-specific languages and language composition; models Dataset, Endpoint, Feature Selection Method, and Classifier; integrates the BDVal plugin to configure BDVal projects; and supports training of models and generation of validation statistics.