Title:Automated Line Tracking of lambda-DNA for Single-Molecule Imaging

Abstract:We describe a straightforward, automated line tracking method to visualize within optical resolution the contour of linear macromolecules as they rearrange shape as a function of time by Brownian diffusion and under external fields such as electrophoresis. Three sequential stages of analysis underpin this method: first, "feature finding" to discriminate signal from noise; second, "line tracking" to approximate those shapes as lines; third, "temporal consistency check" to discriminate reasonable from unreasonable fitted conformations in the time domain. The automated nature of this data analysis makes it straightforward to accumulate vast quantities of data while excluding the unreliable parts of it. We implement the analysis on fluorescence images of lambda-DNA molecules in agarose gel to demonstrate its capability to produce large datasets for subsequent statistical analysis.