PILOT-GM-VAE (Paper)

Patient-Level Analysis of Single Cell Disease Atlas with Optimal Transport of Gaussian Mixtures Variational Autoencoders. We introduce here PatIent-Level Analysis with Optimal Transport based on Gausian Mixture Variational AutoEncoders. PILOT-GM-VAE explores the power of GM-VAE to estimate models describing complex single cell distributions with efficient optimal transport algorithms for estimating the distance between GMs.

Installation

pip install pilotgm

Navigate to Tutorial.

git clone https://github.com/CostaLab/PILOT-GM-VAE.git
cd PILOT-GM-VAE

Then please use the provided Tutorial.

Data sets

You can access the used data sets by PILOT-GM-VAE in Part 1 , Part 2 and Part 3

Citation

@article{joodaki2025pilot,
  title={PILOT-GM-VAE: patient-level analysis of single-cell disease atlas with optimal transport of Gaussian mixture variational autoencoders},
  author={Joodaki, Mehdi and Shaigan, Mina and Samiei, Samaneh and Nagai, James and Mai{\'e}, Tiago and Kuppe, Christoph and Costa, Ivan G},
  journal={Briefings in Bioinformatics},
  volume={26},
  number={5},
  pages={bbaf547},
  year={2025},
  publisher={Oxford University Press}
}