A method, described in Nisslert et al. (2006), for identifying the three-dimensional gel microstructure from statistical information in transmission electron micrographs has been developed in cooperation with SIK (the Swedish Institute for Food and Biotechnology) and Dep. Materials and Surface Chemistry, Chalmers. The micrographs, see the left part of Fig. 3, are projections of stained strands in gel slices.

The gel strand network is modelled as a random graph with nodes and edges, and parameters in the model are estimated by a Markov chain Monte Carlo method. The three-dimensional network may be simulated from the model and the right part of Fig. 3 shows a 3D rendering from such a simulation.

Recently, we have also started a project on simulation of particles diffusing in the estimated gel networks using an adaptive stochastic differential equation solver, after which the estimated diffusion coefficients are compared with diffusion coefficients measured by NMR.

Part of a TEM micrograph from a stained Sepharose gel.

Screen shot from a 3D rendering of the simulated gel network

**Reference **

Nisslert, R., Kvarnström, M., Lorén, N., Nydén, M., and Rudemo, M. (2006) Identification of the three-dimensional gel microstructure from transmission electron micrographs. To appear in Journal of Microscopy.