Johnson's lab develops software (cellPACK) to stochastically pack nanoscale building blocks into cellular ultrastructural data with biologically relevant interactions to generate molecular models of cellular environments. This approach generates probabilistic models at a lower confidence but larger scale than IMP to contribute to a similar goal of assembling unified structural multi-scale models from data spanning diverse disciplines. Currently, Johnson is working with the Sali Lab, Tom Goddard from RBVI, and a larger group to develop a multi-scale hybrid modeling file format and framework to enable more efficient communication and data sharing for the types of models generated by these software packages.

The lab continues to collaborate with Art Olson's Molecular Graphics Lab at The Scripps Research Institute on the ePMV, uPy, and autoPACK projects.