Bayesian inference of agent-based models: a tool for studying kidney branching morphogenesis

• DOI: 10.1007/s00285-018-1208-z
• Publication Date: 2018-02-01T13:17:57Z
• Authors: Ben Lambert; Adam L. MacLean; Alexander G. Flet...; Alexander N. Combes; Melissa H. Little; Helen M. Byrne
• Keywords: Branching (polymer chemistry)

The adult mammalian kidney has a complex, highly-branched collecting duct epithelium that arises as ureteric bud sidebranch from an epithelial tube known the nephric duct. Subsequent branching of to form tree is regulated by subcellular interactions between and population mesenchymal cells surround tips outgrowing branches. produce glial cell-line derived neurotrophic factor (GDNF), binds with RET receptors on surface stimulate several pathways in epithelium. Such are be prerequisite for normal development, although competing theories exist their role morphogenesis. Here we introduce first agent-based model ex vivo uretic branching. Through comparison experimental data, show growth factor-regulated mechanisms can explain early cell branching, but only if division depends switch-like way local concentration; occurring driving level exceeds threshold. We also how recently-developed method, "Approximate Approximate Bayesian Computation", used infer key parameters, reveal dependency parameters controlling factor-dependent switch. These results consistent requirement signals proliferation chemotaxis, both which previously identified roles GDNF.