Recently, I have been attempting to model DNA-ion interactions through a computer program called Delphi. After imputing atomic coordinates, atomic radii, and atomic charges files, Delphi solves the linear and non-linear Poisson-Boltzmann equations and can output the charge neutralized by the ions (can be compared to the charge of the DNA) as well as three-dimensional concentration maps.
Even more recently, I used Matlab to generate a hexagonal array of 19, 60 base pair long, DNA strands (see below). In our experiments the DNA was condensed into a hexagonal array, so these generated atomic coordinates produce a more accurate simulation of the experimental DNA-ion interactions.
Now, I will be working on adjusting parameters in Delphi (grid size, ion valence, ion concentration, boundary conditions, etc.) to match what is expected of the system (almost if not complete charge neutralization of the DNA) to the actual simulation results.