CausalMGM

CausalMGM infers causal relationships from observational biomedical datasets using graphical models to enable causal discovery across multi-modal data including high-throughput sequencing, The Cancer Genome Atlas, and Trans-Omics for Precision Medicine.

Key Features:

• Feature Selection and Clustering: Performs feature selection and clustering to reduce dimensionality and group similar variables in high-dimensional datasets.
• Multi-modal Data Integration: Integrates genetic, genomic, clinical, and behavioral data and accommodates high-throughput sequencing datasets and repositories such as The Cancer Genome Atlas and Trans-Omics for Precision Medicine.
• Automated Causal Identification via Graphical Models: Employs graphical models that represent variables as nodes and edges to identify direct influences and novel associations from observational data.
• Visualization of Learned Causal (Directed) Graphs: Produces visualizations of learned directed graphs to represent inferred dependencies among variables.

Scientific Applications:

• Causal Discovery from Observational Data: Inferring cause-and-effect relationships from observational biomedical datasets to move beyond correlation.
• Systems Biology Interaction Analysis: Characterizing interactions across biological layers (genetic, genomic, clinical, behavioral) in systems biology studies.
• Hypothesis Generation and Experimental Prioritization: Supporting hypothesis generation and the prioritization of potential therapeutic interventions and experimental targets based on inferred causal relationships.
• Analysis of Large Public Repositories: Analyzing multi-modal datasets from The Cancer Genome Atlas, Trans-Omics for Precision Medicine, and other high-throughput sequencing resources.

Methodology:

Performs feature selection and clustering, learns graphical models representing dependencies among variables, and infers directed causal graphs from observational data to identify direct influences and associations.