Fig. 6From: Putative biomarkers for predicting tumor sample purity based on gene expression dataA schematic of the XGBoost workflow. The shaded area indicates the data and its partitioning. The boxes inside the dashed lines depict training and testing procedures where T stands for tree and GBM stands for gradient boosting machine. The two oval boxes on the right denote the outputs from XGBoost. A tree-representation of the training and testing procedures is provided in Additional file 13: Figure S4Back to article page