MVBioDataSim

MVBioDataSim generates synthetic multi-view genomic datasets for benchmarking feature selection and integrative analysis across OMICs such as mRNA expression, miRNA expression, copy number variation, and DNA methylation.

Key Features:

• R implementation: Provided as an R package for dataset generation and analysis integration.
• Multi-view simulation: Produces synthetic datasets covering mRNA expression, miRNA expression, copy number variation, and DNA methylation from identical samples.
• Controlled composition and parameters: Enables specification of dataset composition and predefined parameter values for reproducible simulations.
• Network-based interaction modeling: Creates networks that simulate interactions among biological molecules, particularly regulators of gene expression.
• ODE-based dynamics: Derives synthetic data from ordinary differential equation (ODE)–based models with predefined parameters.
• Biological coherence: Maintains consistency with biological mechanisms to produce realistic synthetic signals suitable for method evaluation.
• Benchmark-ready outputs: Generates datasets with known ground truth for assessing feature selection, validation of molecular interactions, and comparison of data mining techniques.

Scientific Applications:

• Benchmarking feature selection: Provides controlled datasets to evaluate and compare feature selection methods across multi-omics data.
• Method development and validation: Supports development and validation of integrative analysis methods for mRNA, miRNA, CNV, and DNA methylation.
• Assessment of data mining techniques: Enables assessment of clustering, classification, and network inference approaches using realistic synthetic multi-view data.
• Validation of molecular interactions: Supplies ground-truth interaction networks to validate inferred regulatory relationships among genes and regulators.
• Integrative pipeline testing: Allows testing of end-to-end multi-omics analysis pipelines under controlled, known-parameter scenarios.

Methodology:

Networks simulating interactions among biological molecules are created and synthetic datasets are generated by simulating ODE-based models with predefined parameters.