Two- and three-dimensional de-drifting algorithms for fiducially marked image stacks.

Abstract

Traction force microscopy has been established as the accepted method for evaluating cell-induced mechanical stresses to their microenvironments, typically using two-dimensional (2D) elastic, synthetic gel-substrates. As cells naturally experience 3D environments in vivo, traction microscopy has been adapted to 3D gels; cells can be tracked over time in 3D. Microscopy images acquired in several fields-of-view e.g. in a time series, may experience drift, which can produce artefactual results that may appear valid and lead to flawed analysis. Hence, we have developed an algorithm for 2D/3D de-drifting of cell-images on 3D gels with fiducial markers (beads) as anchor points. Both lateral and vertical de-drifting are performed using gel-internalized beads, as those used in traction microscopy experiments; this eliminates need for immobilizing beads under the gel for de-drifting, and reduces experiment time. We introduce simulations of initially grid-ordered dots (beads) that are radially displaced to experimentally observed distances, while also applying additive drift. This facilitates testing and demonstration of the de-drifting procedures in 2D/3D. We demonstrate the importance of applying de-drifting using both computer-simulated drifts and experimentally observed drifts in confocal microscopy images. We show that our de-drifting algorithm can remove lateral and/or vertical drift revealing even small, underlying signals. The 2D/3D de-drifting algorithm, crucial for accurate identification of cell-induced marker-displacement, as well as the bead simulations, will shorten traction microscopy experiments and facilitate optimization of the experimental protocols.