Multiscale Modeling of Myelodysplastic Syndromes

Project: Research project

Project Details


Motivated by the observation of waves of gene induction spreading across the pathway, we plan a novel network analysis method to characterize coordinated expression changes across a pathway. This approach, which is methodologically related to our prior work, employs a novel use of spectral graph theory to summarize gene expression and co-expression measurements in context of the genes’ position in the regulatory network. In contrast to other recent systems-level analyses that require the computation of gene-level statistics before incorporating the network topology, our method summarizes the experimental data at the systems level without relying on single-marker gene-level association tests. This technique is especially well-suited to discovering complex, multi-gene patterns of aberrant signaling in myelodysplastic syndromes and for developing a model of the regulatory "road-map" that governs hematopoietic differentiation dynamics.

Our contributions to this project will be:
1. Carry out the basic biostatistical and bioinformatic analysis of the transcriptomic and proteomic data generated by the Corey lab;
2. Refine and apply our spectral network analysis method to identify the subnetworks of interactions affected by perturbations of GCSFR;
3. Iteratively tune the network models by adding/removing edges, with the goal of identifying new regulatory interactions that may be adaptations to aberrant GCSFR signaling;
4. Investigate how cell-specific GCSFR signaling is integrated through protein kinases to cause cell-specific outcomes (normal homeostasis v. malignant transformation) by developing and applying a "Network-Constrained" Partial Least Squares Regression model to predict multivariate proteomic experimental measurements as a function of the above network characteristics.

This work will provide rigorous bioinformatic analysis of the experimental data, will generate experimentally testable hypotheses about the key genes and interactions governing the SCN -> AML transformations (to be validated in the Corey lab, VCU), and will inform the population dynamics models being developed by the Kimmel group (Rice). It will also establish novel computational tools that could be readily applied to investigate other disease processes.
Effective start/end date4/1/1610/31/18


  • Virginia Commonwealth University (FP00000825_SA003//5R01HL128173-04)
  • National Heart, Lung, and Blood Institute (FP00000825_SA003//5R01HL128173-04)


Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.