PaSCient: Learning multi-cellular representations of single-cell transcriptomics data enables characterization of patient-level disease states

Installation

To install the related packages for model training, please use:

conda env create -f env.yml --name pascient

for creating the environment, and then:

bash relevant_install.sh

for installing the helper packages, and then:

pip install -e .

for installing the target package.

Training

To train the model, use the current path, and then run:

python cellm/scripts/train_classification_model.py fit --config cellm/configs/disease_classifier.yaml --trainer.logger.save_dir <SAVE_DIRECTORY> --trainer.logger.project <PROJECT_NAME> --trainer.logger.entity <ENTITY_NAME>

Application

Please refer the folder application for the experiments we did for disease-state prediction, severity analysis and response prediction.

Please refer the folder reproduce for experiments to reproduce the figures we have in this manuscript.

Contact

If you have any questions, please contact Tianyu Liu (tianyu.liu@yale.edu) or Edward De Brouwer (edward.debrouwer@gmail.com).

Citation

...