Clustering gene expression data using Coxian phase-type survival

The paper is concerned with modeling gene expression data of Acute Myeloid Leukemia patients and their survival. Of particular interest is the development of a clustering approach that will identify clusters of patients that have similar gene expression characteristics. This paper considers one such approach based on the Coxian phase-type distribution which proved to be an adequate representation for the patient survival distribution. If such an approach were introduced it would facilitate better patient management and hence allow the most suitable care plan for the different clusters of patients. This will not only have financial benefit but will provide a better quality of service to the patient who will no longer have to undergo ineffective treatments that may have damaging side effects.