Stochastic modelling and statistical inference of hematopoietic development in blood cancers

Myeloproliferative neoplasm emerges from somatic mutations in hematopoietic stem cells, in particular it is strongly correlated with mutation of gene JAK2 that gives cells a competitive advantage over wild-type cells. Understanding the dynamics of mutated cell populations and disease progression could enhance diagnostics and treatment. Due to the impracticality of directly observe hematopoietic stem cell behavior in humans, a mathematical model for cell population development is proposed. This model must capture biological processes' inherent variability and stochasticity, with parameters inferred from real data. This internship builds on previous work modeling the development of JAK2-V617F mutated stem cells, using Approximate Bayesian Computation for parameter inference from patient data. The model's extension involves applying it to data from additional patients and further refining the inference algorithm, in particular concerning the definition of distance that approximates the likelihood in a standard Bayesian framework.