Abstract
Even under the same treatment, responses can vary. Virtual cohorts build on an available, often limited, dataset can capture these differences and enable the discovery of treatment protocols that work well for a wide variety of individuals. In this paper, we refined current virtual cohort pipelines by improving data handling, ensuring the virtual cohort can be used to stratify individuals into treatment subgroups based on their data, and validating that the virtual cohort matches the observed data variability. To illustrate, we applied this pipeline to a murine data set of orthotopic bladder cancer treated with gemcitabine (Gem) and immunotherapy with OT-1 cells. We generated over 10,000 virtual mice that replicate the dynamics of three cell subpopulations in the tumor (cancer cells, T cells, and myeloid-derived suppressor cells) and data from four experimental cohorts (control, Gem, OT-1, and Gem+OT-1). We also provided a guide for using this pipeline for other treatments.
Full Text
The Full Text of this preprint is available as a PDF (1.6 MB). The Web version will be available soon.