PETIL: Predicting Expansion of Tumor Infiltrating Lymphocytes for the Adoptive Cell Immunotherapy in Bladder Cancers

One major advance in treating solid tumors is the success of adoptive cell therapy (ACT), in which autologous tumor-infiltrating lymphocytes (TILs) are expanded and activated ex vivo and then reinfused into the cancer patient.

PETIL is a tool that can first learn from patient and tumor data already collected in the clinic (local data) which data features are important for predicting TIL expansion, without the need to predefine which data categories to consider. Then, this tool predicts a possible TIL expansion for individual patients (personalized predictions), allowing to determine whether ACTTIL therapy could potentially treat an individual bladder cancer patient.

Repository structure

.
├── 01_Data_summary.ipynb          # Exploratory analysis and dataset summary
├── 02_Data_table.ipynb            # Generate descriptive table
├── 03_norm.ipynb                  # Data preprocessing and normalization
├── 04a_midas_train.ipynb          # MIDAS imputation on training cohort
├── 04b_midas_test.ipynb           # MIDAS imputation on testing cohort
├── 04c_Imputation_analysis.ipynb  # Evaluate imputation performance
├── 05_MI.ipynb                    # Mutual Information feature ranking
├── 06_p_corr.ipynb                # Pairwise correlation analysis
├── 07a_FFS.ipynb                  # Forward Feature Selection
├── 07b_adeq_smplSz.ipynb          # Sample size adequacy analysis
├── 07c_Stab_FeatSelection.ipynb   # Feature selection stability analysis
├── 08_MCC_RBF_SVM.ipynb           # RBF-SVM model with MCC threshold optimization
│
├── data/                          # Original and processed datasets
├── figs/                          # Figures generated by notebooks
├── utility/                       # Utility functions used throughout the project
├── midas/                         # MIDAS imputation implementation
├── smpl_sz_adeq/                  # Sample size adequacy module
├── midas_train/                   # Imputed training datasets
├── midas_test/                    # Imputed testing datasets
├── tmp/                           # MIDAS checkpoints
│
├── py_libraries.py                # Python library imports
├── README.md                      # Project documentation
└── Feature_Distri.ipynb           # Distribution plots for selected features

Authors

Kayode Olumoyin [email protected], Katarzyna Rejniak

Source Code

https://github.com/okayode/Predictor_of_the_Expansion_of_TIL_project

License

This project is licensed under the GNU General Public License v3.0.