Introduction to causality and graphical models in biology

M2 GENIOMHE course (Université d’Évry): causal graphs, d-separation, biological networks, interventions, Mendelian randomisation and causal discovery with LiNGAM.

This course, taught in the M2 GENIOMHE at Université d’Évry, introduces causal reasoning with graphical models for bioinformaticians. It starts from directed acyclic graphs and the factorisation they encode, moves to the networks used in biology (Gaussian graphical models, the PC algorithm), then to interventions, Mendelian randomisation and the discovery of causal orders with LiNGAM.

The slides are available in two forms: interactive HTML slides and a printable PDF version, with four slides per page.

Illustration linking a causal network of cells, an intervention that cuts the arrow into a variable, and a non-Gaussian point cloud with a causal direction

Lectures

Session	Topic	HTML	PDF
1	Causality, DAGs, d-separation	Slides	Printable version
2	Biological networks: Gaussian graphical models, PC algorithm	Slides	Printable version
3	Interventions, Mendelian randomisation, LiNGAM	Slides	Printable version

Tutorials

Tutorial booklet — PDF version without solutions: theory exercises (pen and paper) and Python practice for the three sessions.