knowledge graph facilitated drug response prediction

Data processing

Generate feature matrix for the training set Beat AML Wave 1/2
Generate feature matrix for the test data set Beat AML Wave 3/4
Generate feature matrix for the Functional Precision Medicine Tumor Board

Construct a minimal graph to facilitate drug response prediction for Acute Myeloid Leukemia (AML)

Extract TF miRNA to target gene pairs for the knowledge graph
Fine house keeping genes as reference

Building predictive models

XGBoost Classifiers for Venetoclax using different feature sets

XGBoost Classifiers for Trametinib using different feature sets

XGBoost Classifiers for all drugs using different feature sets

Feature exploration

Feature annotation

Regression models

Regression analysis Regression result visulization