This is the comprehensive source code for my Master's Thesis on Drug Response Prediction using Gene Expression and Molecular Structure
Author:
- Sun Yih-Yun (孫懿筠)
- data
Including both the raw and preprocessed data along with the preprocessing process in this work - Computational_method_comparison
Comparing the performance of different models for the task of drug response prediction (drug-blind testing: unknown compounds and known cell lines)- Matrix-Factorization model (MF_model)
- Machine Learning model (ML_model)
- NN model with SMILESVec protein representation (SMILESVec_model)
- First-order Weisfeiler-Lehman GNN model (WL_GNN_model)
- Combined_model_evaluation
Constructing a combined (2-step) model for the prediction of new drugs- CaDRReS_CLsim: Drug response prediction on known compounds and unknown cell lines based on the similarity of cell lines (prediction of cell-blind testing set)
- CaDRReS_CLsim_SVM: Drug response prediction on unknown compounds and unknown cell lines based on the similarity of molecular structure (prediction of disjoint testing set)
- Model_comparison
Comparing the performance of the combined model with Precily on external testing data - User-friendly_interface
Developing a user-friendly drug response prediction tool using a Docker image - Discussion
- lineage_analysis
- dose_range
- parameter_tuning
- System apps
- Python 3.10.0+
- Python packages
- requirements.txt