TERMs are tertiary motifs that comprise protein structures. In Zheng et al.[1] we suggest that PDB-based TERM statistics represent fundamental relationships between sequence and structure. We validate this by demonstrating that breaking structural models into their constituent TERMS, and mining the PDB for close matches to each, enables one can differentiate good from poor models, and even identify poorly predicted regions within models [1].
The code for performing this analysis, TERMANAL, is freely available to academic users under the terms of the GNU General Public License and can be downloaded from here. Note that the download includes also all of the support data we used in the paper, and in particular the specific structural database that was used. Thus, if you run the code as is, it will correspond exactly to the procedure performed in the paper. But you can also easily use a different structural database that may be better for your application or is just more up to date; see the README file for instructions.
If you use TERMANAL in your research, please cite the following paper:
[1] F. Zheng, J. Zhang, G. Grigoryan, "Tertiary Structural Propensities Reveal Fundamental Sequence/Structure Relationships", Structure, 23(5): 961-971, 2015.