Joint Parameter Estimation and Base-Calling for Pyrosequencing Systems

Sequencing-by-synthesis technology takes us a step closer to the promises of personalized medicine by enabling affordable and fast DNA sequencing. Its accuracy, however, is fundamentally limited by the imperfections of the underlying biochemical processes and signal acquisition noise. In this paper, we focus on pyrosequencing systems, derive a mathematical model of the signal they acquire, and develop a joint parameter estimation and base-calling procedure which relies on Markov chain Monte Carlo (MCMC) and iterative least squares techniques. Simulations and experimental results demonstrate that the MCMC algorithm outperforms existing state-of-the-art base-calling method.