CausalTrail

CausalTrail performs causal hypothesis testing using Causal Bayesian Networks and do-calculus to model and quantify effects of interventions in biological systems.

Key Features:

• Causal Bayesian Networks: Uses Causal Bayesian Networks to encode causal relationships between variables for intervention analysis.
• Do-Calculus Integration: Applies do-calculus to reason about and compute the effects of hypothetical interventions.
• Intervention Simulation: Supports simulation and analysis of interventions such as gene knockouts and drug administrations to evaluate downstream effects.
• Multiple Data Import Methods: Accepts diverse input datasets through multiple import methods for integration of heterogeneous bioinformatics data.
• Flexible Query Language: Provides a query language to formulate and evaluate complex causal hypotheses on the network.
• Handling Missing Data: Incorporates procedures to account for missing data, enabling analysis in multi-omics and incomplete datasets.
• Downstream Analysis of CBNs: Enables downstream computational analyses on constructed Causal Bayesian Networks to support hypothesis testing.

Scientific Applications:

• Gene Expression Analysis: Models causal relationships to assess how interventions affect gene expression patterns.
• Drug Response Prediction: Evaluates potential causal effects of drug administrations on biological outcomes for response prediction.
• Systems Biology Causal Inference: Supports causal inference in systems biology to elucidate mechanisms and intervention impacts.

Methodology:

Implemented in C++ using the Boost and Qt5 libraries.