Research Article: The ‘Densitometric Image Analysis Software’ and Its Application to Determine Stepwise Equilibrium Constants from Electrophoretic Mobility Shift Assays

Date Published: January 21, 2014

Publisher: Public Library of Science

Author(s): Liesbeth van Oeffelen, Eveline Peeters, Phu Nguyen Le Minh, Daniël Charlier, Mark R. Muldoon.


Current software applications for densitometric analysis, such as ImageJ, QuantityOne (BioRad) and the Intelligent or Advanced Quantifier (Bio Image) do not allow to take the non-linearity of autoradiographic films into account during calibration. As a consequence, quantification of autoradiographs is often regarded as problematic, and phosphorimaging is the preferred alternative. However, the non-linear behaviour of autoradiographs can be described mathematically, so it can be accounted for. Therefore, the ‘Densitometric Image Analysis Software’ has been developed, which allows to quantify electrophoretic bands in autoradiographs, as well as in gels and phosphorimages, while providing optimized band selection support to the user. Moreover, the program can determine protein-DNA binding constants from Electrophoretic Mobility Shift Assays (EMSAs). For this purpose, the software calculates a chosen stepwise equilibrium constant for each migration lane within the EMSA, and estimates the errors due to non-uniformity of the background noise, smear caused by complex dissociation or denaturation of double-stranded DNA, and technical errors such as pipetting inaccuracies. Thereby, the program helps the user to optimize experimental parameters and to choose the best lanes for estimating an average equilibrium constant. This process can reduce the inaccuracy of equilibrium constants from the usual factor of 2 to about 20%, which is particularly useful when determining position weight matrices and cooperative binding constants to predict genomic binding sites. The MATLAB source code, platform-dependent software and installation instructions are available via the website

Partial Text

To calibrate an autoradiograph, a calibrated step tablet should be scanned, preferrably together with the autoradiographic film to be analyzed, as the sensitivity of the scanner may change over time. Based on the relation between the average gray values of the steps in the step tablet and their optical densities, the measured gray values in the autoradiograph can be mapped to optical densities by curve-fitting or linear interpolation. However, the relation between optical density and concentration of radioactive material is not linear, and therefore, a second mapping should be performed, from optical densities to relative concentrations (1)with and the minimum and maximum OD’s of the autoradiographic film, i.e., the OD after development of an unexposed film and a film exposed to broad daylight [1]. Current software applications for densitometric analysis, such as ImageJ, QuantityOne (BioRad) and the Intelligent or Advanced Quantifier (Bio Image) do not offer the possibility of a second mapping and/or a fit to the curve described by equation 1, and therefore display the same non-linearity as the film. The ‘Densitometric Image Analysis Software’ has been developed to solve this issue. Moreover, apart from measuring electrophoretic band intensities in autoradiographs, gels and phosphorimages, this software also allows to determine binding constants from Electrophoretic Mobility Shift Assays (EMSAs).

This section describes the entire program, how cooperative binding constants and their standard deviations can be determined, and a method to simulate smear caused by the dissociation of protein-DNA complexes.

In this section, we discuss the possibilities to deal with the different error sources observed when determining stepwise equilibrium constants, and validate our method.