In our research, we are dealing with the problem of computationally identifying the domains of large macro-molecules (such as Chaperonin) that remain conserved upon conformation change, by comparing low-resolution volumetric electron microscopy (EM) maps. A simple instance of this problem can be shown to be NP-complete. As of now there is no computational tool to solve this problem efficiently or even approximately. We are developing the first method to do so.

Previous attempts have been for some simple instances of this problem, like identifying secondary structures in intermediate-resolution maps (helixhunter: http://ncmi.bcm.tmc.edu/software/AIRS/ssehunter/sse-help.htm) and docking a component of a bio-molecular assembly into latter's EM map. For intermediate resolution maps (10A), where secondary structures "fuzz-out" beyond the recognition limits of helixhunter or even any manual technique. However, we believe that it should be still possible to extract motifs or domains which are bigger entities than secondary structures and could remain in the "detectable range" may be upto 20A resolution.