R.ROSETTA

R.ROSETTA constructs interpretable rule-based classification models to analyze decision-related omics and transcriptomic data and to reveal co-predictive feature relationships relevant to biological processes, including applications in case-control studies of autism and neurodevelopmental genes.

Key Features:

• Rule-Based Modelling: R.ROSETTA employs rule-based modeling to derive combinatorial statistics that capture co-predictive feature relationships.
• Interpretability and Transparency: It produces interpretable, transparent classification models that expose rule semantics across diverse data types.
• Bias and Noise Minimization: The framework provides statistical and visualization methods to assess and reduce analysis bias and noise.
• Versatility Across Omics Data: It handles decision-related omics data and has been demonstrated on case-control transcriptomic studies of autism.

Scientific Applications:

• Feature Interdependency Analysis: Investigating interdependencies among features that distinguish phenotype classes to identify co-predictive mechanisms.
• Neurodevelopmental and Autism Gene Studies: Generating hypotheses about autism-related and other neurodevelopmental gene associations from transcriptomic case-control data.

Methodology:

Constructing rule-based classification models, computing combinatorial statistics, and applying statistical and visualization methods to identify rule semantics and co-predictive mechanisms.