Professor Geissler and graduate student David Sivak developed coarse-grained nucleic acid models to characterize the bending elasticity of double-stranded (ds) DNA at short lengthscales. We worked closely with Professor Jan Liphardt's laboratory to understand the biophysical implications of their experiments using FRET to measure internal forces generated in small loops composed of partly double-stranded, partly single-stranded DNA. More recently we collaborated with Professor Paul Alivisatos's laboratory to provide mechanistic understanding of their experiments, which scatter X-rays off gold nanoparticles attached to the ends of short dsDNA strands, allowing the measurement of the entire equilibrium ensemble of dsDNA conformational fluctuations, free in solution without external perturbation. In each set of experiments, we used standard polymer models for chain bending elasticity, careful modeling of boundary conditions between dsDNA and alkane linkers, and Monte Carlo importance sampling, we calculate probability distributions for distances between the gold nanoparticles, seen above for 6 different lengths of intervening dsDNA. Our calculations showed that 50-nm persistence-length wormlike chain (WLC), a successful model at longer lengths (kilobases of DNA) in single-molecule force-extension experiment, also is consistent with the data at the much shorter lengths (tens of bases of DNA) of these experiments. We also found that these experimental ensembles are not sensitive to details that make large differences in cyclization experiments, such as the presence and activation free-energy of thermally-excited 'melts' with greatly-increased flexibility. We also demonstrated that similar models are useful for calculating cyclization efficiencies of dsDNA strands too short for other computational methods, and for understanding the kinetic barriers involved in dsDNA threading through nanopores. We are currently exploring the agreement of these coarse-grained models of dsDNA bending with statistics of all-atom molecular dynamics simulations.