Consensus Enhances Individual Causal Models: a Use Case on Lung Cancer Driven by Cellular Pathways
Published in 24th International Conference on Computational Methods in Systems Biology (CMSB 2026), 2026
Discovering reliable cause-and-effect relationships in real-world medical data remains an open challenge. Classical Causal Discovery (CD) algorithms rely on strict assumptions — functional form, data distribution, causal sufficiency — that are rarely met in practice, compromising result interpretability and downstream decision-making. To address this, we introduce a consensus causal model that combines multiple CD algorithms to enhance accuracy. Our model is constructed from heterogeneous causal graphs via a homogenisation step that ensures semantic compatibility and enables meaningful information exchange. Applied to a lung cancer dataset linking smoking, age, tumor stage, and cellular pathway mutations, we find that individual CD algorithms produce significant structural inconsistencies, while the consensus model effectively aggregates their strengths and mitigates their uncertainties. The resulting model reveals biologically validated causal relationships that isolated algorithms fail to capture, underscoring consensus causal modelling as a robust alternative to single-model selection.
Recommended citation: Lang A., Henrique Ramos R., Al-Ali S., Reza Mousavi M., Calissano A., Balelli I. (2026). “Consensus Enhances Individual Causal Models: a Use Case on Lung Cancer Driven by Cellular Pathways.” 24th International Conference on Computational Methods in Systems Biology (CMSB 2026).