"Probing immune cell wound recruitment signals using Bayesian inference and random walk models"
The recruitment of immune cells to wounds is a complex spatiotemporal process with the production and diffusion of chemoattractants acting as a beacon for immune cells to respond to damage within the body. Analysing these chemoattractants can be experimentally complex, however, inference of the chemoattractant field is possible by analysing cell trajectories. These trajectories can then be used to infer the main parameters of the underlying chemoattractant.
To undertake this study, we reproduced a previously published modelling framework which utilises biased-persistent random walk to capture immune cell motion and the diffusion equation to capture the chemoattractant dynamics. By applying Bayesian inference, this framework allows us to gain an understanding of the relationship between cell migration parameters and the main chemoattractant parameters such as the diffusion coefficient and production time.
To aid transparency, we implemented an open source version of the modelling framework to allow for future research. We then applied this model to investigate the chemoattractant which is responsible for wound healing within Drosophila, this chemoattract is currently unknown and can be difficult to isolate through experiments. However, by applying the inference model it is possible to isolate the gene responsible for the expression of the chemoattractant. We compared wild type and gene deleted (mutant) datasets and found a significant difference between inferred parameters, which implies that gene deletion is consistent with no production of chemoattractant.