Research Paper Volume 15, Issue 12 pp 5826—5853

Identification and validation of a 7-genes prognostic signature for adult acute myeloid leukemia based on aging-related genes


Figure 1. Modelling of aging-related genes (ARGs) prognostic signature. (A) Random survival forest algorithm for selecting hub ARGs. (B) The raw importance of the 32 candidate ARGs. (C) Weighted random forest (Ranger) was preformed to choose the optimization model with a lowest out of bag (OOB) error rate. (D) Seven ARGs with a relative importance greater than 0.52 were transferred to construct prognostic signature using COX regression model.