Transposable Element Expression as a Molecular Linker of Transcriptomic States Associated with Induction Failure in Acute Myeloid Leukemia
Scripts and data used in our paper titled "Transposable Element Expression as a Molecular Linker of Transcriptomic States Associated with Induction Failure in Acute Myeloid Leukemia"
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.
Figure 1: Contains data and scripts used to produce Figure 1
Figure 2: Contains data and scripts used to produce Figure 2
Modelling: Contains data and scripts used in the XGBoost models