Mila > Learning Causal Models in Single Cell Dynamics – A Mila-Helmholtz International Project

Combining AI Expertise to Tackle Cellular Processes in Disease

Keywords: Mila, Helmholtz, MPI, AI, Bioinformatics, Deep learning, Causal learning, Cellular dynamics, Single cell genomics, Computational genomics, Machine learning, Yoshua Bengio


Mila and its partners are committed to helping researchers in stem cell biology, biomedicine and related fields respond to some of the world’s most urgent health challenges such as cancer, diabetes, and antimicrobial resistance.

As part of this mission, Mila has joined forces with the research institute Helmholtz Munich and the Max Planck Institute for Intelligent Systems to form an international lab that will use big data and artificial intelligence to shed light on the role of cellular processes in the development of diseases. 

Project Description

Recent advances in genomics allow researchers to measure individual cellular differences with unprecedented resolution and see more clearly how cell characteristics are influenced by their environment. These breakthroughs have generated a wealth of molecular biology data for which the analysis and interpretation are of critical importance.

To address this challenge, Mila and its partners recently launched the Causal Cell Dynamics Lab. The initiative merges Mila’s deep learning expertise, Helmholtz’ genomics-specific bioinformatics skills and causal learning specialists from the Max Planck Institute. The Causal Cell Dynamics Lab researchers have a five-year mandate to find ways to identify underlying causal structures that explain cell development and  trajectories and to develop models that improve our of cell responses to environmental changes. The Lab is being co-funded by the Helmholtz Initiative and Networking Fund.

The Canada-Germany collaboration will forge strategic partnerships between leading-edge research institutions, allowing world-renowned researchers and promising PhD students to work abroad and strengthen international connections and networks.

Project Aims

The goal of this research effort is the modeling of cellular dynamics from single cell genomics as well as the latent causes driving those dynamics. Through our work, we are developing methods for causally structured deep representation learning to improve understanding of cellular decisions. 

Specifically, our aims are to:

  • Develop robust high-dimensional causal inference working on latent spaces
  • Develop causal models for cell dynamics and perturbations
  • Extend to multi-omics data adding biological priors


Mila and Helmholtz Munich combine their expertise, forming an International Lab

The Human Cell Atlas is an international collaborative consortium that charts the cell types in the healthy body, across time from development to adulthood, and eventually to old age.


The Causal Cell Dynamics Lab is a collaboration between Helmholtz Munich, the Max Planck Institute for Intelligent Systems and Mila.

Lab Directors

Fabian Theis (Helmholtz)

Other Members

Joseph Paul Cohen (Mila)

Benjamin Schubert (Mila)

Alexander Tong (Mila)

Maria Colomé-Tatché (Helmholtz)

Niki Kilbertus (Helmholtz)

Annalisa Marsico (Helmholtz)


For any inquiry about the project, please contact Mathieu Bourgey, Scientific Manager (Mila) at or Mathieu Seyfrid, Program Manager (Helmholtz) at