STILT - Stochastic Inference on Lineage Trees
STILT is a software package for performing exact Bayesian parameter inference and model selection on the basis of quantified time-lapse fluorescence microscopy movies of populations of proliferating cells. STILT uses a particle filter-based approach to infer the latent history of unobserved biochemical species and the posterior distribution of biochemical parameters, for user-specified mechanistic, stochastic chemical reaction/gene regulation models. It estimates the marginal likelihood (evidence) of the specified models, from which one can rank candidate models and perform model selection using Bayes Factors. Using a goodness-of-fit test, candidate models and inferred model parameters can be compared for plausibility in the context of measured data.
The STILT software (including example data) is available on Github at: https://github.com/claassengroup/STILT/