{"references": ["Dundar M, Yerebakan HZ, Rajwa B. Batch discovery of recurring rare classes toward identifying anomalous samples. In: Proc. 20th ACM SIGKDD Int. Conf. Knowl. Discov. Data Min. KDD '14. New York, NY, USA: ACM; 2014. p 223\u2013232. Available at: http://doi.acm.org/10.1145/2623330.2623695", "Dundar M, Akova F, Yerebakan HZ, Rajwa B. A non-parametric Bayesian model for joint cell clustering and cluster matching: identification of anomalous sample phenotypes with random effects. BMC Bioinformatics. 2014;15:314. Available at: https://doi.org/10.1186/1471-2105-15-314", "Rajwa B, Wallace PK, Griffiths EA, Dundar M. Automated assessment of disease progression in acute myeloid leukemia by probabilistic analysis of flow cytometry data. IEEE Trans Biomed Eng. 2017;64(5):1089\u201398. Available at: https://doi.org/10.1109/TBME.2016.2590950", "Tario JD, Wallace PK. Reagents and cell staining for immunophenotyping by flow cytometry. In: McManus LM, Mitchell RN, editors. Pathobiology of Human Disease. San Diego: Academic Press; 2014. p. 3678\u2013701. Available from: https://doi.org/10.1016/B978-0-12-386456-7.07104-5"]}

Following the endorsement of the Bethesda International Consensus Group, as well as the guidance for classification of hematolymphoid neoplasms introduced by the WHO's Classification of Tumours of Haematopoietic and Lymphoid Tissues, flow cytometry (FC) immunophenotyping became a standard tool for Acute Myeloid Leukemia (AML) diagnosis and disease monitoring. The goal of the presented work is building a functional prototype of an automated non-parametric Bayesian clinical decision–support system that can not only recognize the difference between normal and abnormal samples, but most importantly also recognize the direction of change in disease progression on the basis of the FC bone-marrow data alone, without the need to supplement the data with morphology and/or genetic analysis.