Bayesian Analysis of Immune Response Dynamics with Sparse Time Series Data

Fernando V Bonassi, Cliburn Chan, Mike West
Google, Duke University Medical Center, Duke University

Apr 30 2014

In vaccine  development, the temporal  profiles of relative  abundance of  subtypes of immune cells (T-cells) is key to understanding vaccine efficacy.   Complex and expensive experimental studies generate very sparse time series data on this immune response. Fitting multi-parameter dynamic models of the immune response dynamics-- central to evaluating  mechanisms underlying vaccine efficacy-- is challenged by data sparsity.  The research reported here addresses this challenge.  For HIV/SIV vaccine studies in macaques, we:  (a) introduce novel dynamic models of progression of cellular populations over time with relevant, time-delayed components reflecting the vaccine response; (b)  define an effective Bayesian model fitting strategy that couples Markov chain Monte Carlo (MCMC) with Approximate Bayesian Computation (ABC)-- building on the complementary strengths of the two approaches, neither of which is effective alone; (c) explore questions of information content in the sparse time series for each of the model parameters, linking into experimental design and model simplification for future experiments;  and (d) develop, apply and compare the analysis with samples from a recent HIV/SIV  experiment, with novel insights and conclusions about the progressive response to the vaccine, and how this varies across subjects.


Approximate Bayesian Computation (ABC), Bayesian Inference, Dynamic Models, Immunology, Learnability of Parameters, Markov chain Monte Carlo (MCMC), ODE Models, Sparse Data, Time Delays, Vaccine Design


PDF icon 2014-02.pdf

BibTeX Citation: 

  author = {F. V. Bonassi and C. Chan and M. West},
  title = {Bayesian analysis of immune response dynamics with sparse time series
  institution = {Department of Statistical Science, Duke University},
  year = {2014},
  type = {Discussion Paper},
  note = {Submitted for publication},
  url = {}