Integrative Multi-Layer Modeling Identifies an Endogenous Retrovirus Signature of Induction Failure in AML
Scripts and data used in our paper titled "Integrative Multi-Layer Modeling Identifies an Endogenous Retrovirus Signature of Induction Failure in AML"
Nested gradient-boosted decision tree models were implemented to evaluate incremental predictive value across biological layers for patients receiving standard 3+7 induction. Patients were divided into discovery (BEAT-AML Waves 1-2; n=186) and independent validation (Waves 3-4; n=85) cohorts to ensure temporal robustness and train the XGBoost classifier.
Original_submission: Data and scripts used in the original submission
|_Figure 1: Data and scripts used to produce Figure 1
|_Figure 2: Data and scripts used to produce Figure 2
|_Modelling: Data and scripts used in the XGBoost models
Revision: Data and scripts used in the Revision
|_Revised_Figure2: Data and scripts used to produce Figure 2 in the revision
|_Supplementary_Figure_1: Data and scripts used to produce Supp. Figure 1 in the revision
|_Supplementary_Figure_2: Data and scripts used to produce Supp. Figure 2 in the revision
|_Table_S6: Data and scripts used to produce Table_S6 in the revision
|_Train_vs_Test_Performance: Scripts used in assessing model performance