Clustering and classification of crystal structures is hot. Parkin hit the front cover\nof CrystEngComm with a story on Comparing entire crystal structures: structural genetic fingerprinting\n(DOI:10.1039/b704177b). Now, the story itself, while rather interesting and well written, has three major flaws:
\n\nWell, the latter sounds a bit boohoo, and it is :) (BTW, I do like this paper.)
\n\nThey propose the work as proof-of-concept, but use a very artificial data set of only 12 crystal structures (benzene\nand eleven polycyclic aromatic hydrocarbons, like\nnaphtalene, anthracene,\nphenanthrene, triphenylene,\npyrene, perylene, and coronene).\nWhile such a small set does make a nice example where you can still list all similarities (0.5*N*(N-1)), it is really too artificial.
Now, you may wonder if I am in the position to criticize this shortcoming, but I think I am. As part of my PhD\nwork, I analyzed this problem myself, and published two years ago the paper Method for the computational comparison\nof crystal structures (DOI:10.1107/S0108768104028344). Apparently,\nParkin was not aware of this publication and did not cite it. I should have went to a crystallography conference\nwith a poster, and advertise my work more. In this paper, I analyzed a data set with 48 crystal structures, manually\nvalidated by visual inspection, resulting in having to compare 1128! crystal structure pairs. Took me two full weeks\nbehind a Silicon Graphics. Yes, I really understand why they took only 12 structures :)
\n\nHowever, there is more prior art. While my approach was based on a new radial distibution function-based whole\ncrystal structure descriptor, my supervisor (Ron) used\nthe more common powder diffraction pattern and showed in Representing Structural Databases in a Self-Organising Map\n(DOI:10.1107/S0108768105020331) it to be a good enough descriptor for\nclustering of thousands of crystal structures using a self-organizing map\n(SOM).
\n\nLast week, my second paper in crystallography appeared: Supervised Self-Organizing Maps in Crystal Property and\nStructure Prediction (DOI:10.1021/cg060872y). In this paper, we show how\nsupervised SOMs (see DOI:10.1016/j.chemolab.2006.02.003) can be\nused for supervised classification and even for property prediction. Note that these supervised SOMs are truly\nsupervised, unlike many earlier modifications of the unsupervised SOMs: the training is supervised.
\n\nFinally, another advantage of this last work: the code is open source. The code for the unsupervised SOMs is available as\nR package: kohonen; and for\npowder diffraction patterns: wccsom. Details can be found in\nthis R News issue. The first package is not actually limited to\ncrystal structures, and can be used for any clustering problem. However, the articles mentioned here make use of simulated\ndiffraction patters, and I am not sure there are open source tools to generate those.
\n\nBTW, I would still be interested in teaming up with CrystalEye in\none way or another, and couple these data analysis methods to live streams of new crystal structures. Nick, let me\nknow if you are interesting in idea exchange.
\n\nGetting back to Parkin’s paper, I do like the work. Hirshfield surfaces are an interesting tool to visualize packing\ncharacteristics, and using them to describe a crystal structure sounds like an interesting idea indeed. I just hope\nthat the method properly scales.
\n\n