Author(s) :

Abhigya Bandugula.

Volume/Issue :

Abstract :

Acute Myeloid Leukemia (AML) is associated with a high relapse rate due to the persistence of leukemia cell clones. Risk prediction of relapses can be helpful to create specialized therapies that increase chances of long-term survival for those at risk, especially pediatric patients. Therefore, genomic-expression patterns that are useful in predicting risk of recurrence in patients were focused on for analysis. Using RNA-seq data from the GDC’s TARGET-AML cohort, various feature selection, dimensionality reduction, machine learning, and differential expression analysis methods were applied to find the differences in expression. Mitochondrial and ribosomal protein genes were found to have the largest variation. Genes upregulated in primary patients were involved in cellular respiration and ATP production and genes upregulated in recurrent patients were involved in DNA organization. These results highlight the underlying mechanisms behind primary and recurrent AML and provide insight on further risk assessment and therapy methods.