Closed-Form Formulas vs. PDE Based Numerical Solution for the FRAP Data Processing: Theoretical and Practical Comparison

Abstrakt

Fluorescence recovery after photobleaching (FRAP) is a widely used method to analyze (usually using fluorescence microscopy) the mobility of either fluorescently tagged or autofluorescent (e.g., photosynthetic) proteins in living cells. The FRAP method resides in imaging the recovery of fluorescence intensity over time in a region of interest previously bleached by a high-intensity laser pulse. While the basic principles of FRAP are simple and the experimental setup is usually fixed, quantitative FRAP data analysis is not well developed. Different models and numerical procedures are used for the underlying model parameter estimation without knowledge of how robust the methods are, i.e., the parameter inference step is not currently well established. In this paper we rigorously formulate the inverse problem of model parameter estimation (including the sensitivity analysis), making possible the comparison of different FRAP parameter inference methods. Then, in a study on simulated data, we focus on how three different methods for inference influence the error in parameter estimation. We demonstrate both theoretically and empirically that our new method based on a solution of a general initial-boundary value problem for the Fick diffusion partial differential equation exhibits less bias and narrower confidence intervals of the estimated diffusion parameter, than two closed formula methods.