ICDD is rumored not to be amused by the contribution of the HighScore-based search functionality, and rumored to be claiming breach of intellectual property. I have not seen either any ICDD patents nor the HighScore implementation, but clearly there is a conflict of interest.

BTW, Panton Principle endorsers may be interesting in signing the petition for Open Data in crystallograph too.

Instead of setting up a BTS at SourceForge, locally with Bugzilla, or at LaunchPad, he suggested to use Connotea:

To report a problem in CrystalEye, simply bookmark an example of the problem with the tag “crystaleyeproblem”, using the Description field to describe the problem. All the problems will appear on the tag feed.

When we fix the problem we’ll add the tag “crystaleyefixed” to the same bookmark. If you subscribe to this feed, you’ll know to remove the crystaleyeproblem tag.

In the fullness of time, we’re planning to use connotea tags to annotate structures where full processing hasn’t been possible (uncalculatable bond orders, charges etc).

Now, Connotea is advertised as a [f]ree online reference management for all researchers, clinicians and scientists, and I have never really been happy with any HTML page ending up in the system, I would counter the suggestion by using social bookmarking websites for any HTML page (not just publications), such as Del.icio.us (see their list of CrystalEye bookmarks).

Anyway, it does not really matter, and Connotea has an open API to query the database. This will allow Jim to write a simple userscript to enhance each CrystalEye page with a list of bug reports. That will allow every CrystalEye visitor to see what others are commenting on it. In that respect, many other things can be envisioned… Getting comments on the paper behind the crystal structure from Chemical blogspace and Postgenomic, …

1. number of bounded neighbors
2. number of pi bonds (or double bond equivalents)
3. number of lone pairs
4. hybridization state

This seems to be a minimal and accurate set to cover a rather good deal of chemoinformatics. I have yet to make the mappings of the new atom type list with existing lists for force fields, and radicals are missing too. However, the following algorithms in the CDK seem to translate rather well:

• hydrogen adding
• aromaticity detection (Hückel rules)

I still have to rework the double bond perception.

Aromaticity

Now, aromaticity is a fuzzy concept, and there is no general agreement on what it is. Some say it is smelly compounds, others say ring systems which apply to the Hückel rule. Based on the new atom type list, I have rewritten the Hückel aromaticity detector and it applies these rules:

• only single rings and two fused non-spiro rings
• 4n+2 electrons
• no ring atoms with double points not in the ring too

This approach differs in two ways from the old code: it no longer tries to test all ring systems, which required to use the CDK AllRingsFinder algorithm which combinatorial generates all possible ring systems. The new code only considers ring systems with up to two single rings. Aromaticity beyond that is even less well defined than aromaticity in general.

The other difference is that the ring system must not have ring atoms which have a double bond which is not part of the ring too. The classical example is benzoquinone (InChI=1/C6H4O2/c7-5-1-2-6(8)4-3-5/h1-4H) which is not aromatic, even though it conforms the 4n+2 rule (image from PubChem):

Evidence of Aromaticity

The final rule, of course, is what nature tells us what is aromatic and what is not. There are many other details to aromaticity than I just covered. For example, take azulene (InChI=1/C10H8/c1-2-5-9-7-4-8-10(9)6-3-1/h1-8H). All atoms are aromatic, but not all bonds (also PubChem):

These things are complex, but the rise of Open Data helps us out, as well as increasing computing power. Peter has been running two rather projects which may help us out: CrystalEye (Nick: no blog?) and OpenNMR .

NMR shifts will give us experimental backup on our notion of aromaticity, and so do bond lengths. I asked Peter about this , and whether OpenNMR predicted shifts could indeed confirm aromaticity of compounds, and he replied and showed that the predicted spectra could be used to distinguish between C-C and C=C bonds.

I commented the following (which was in moderation at the time of writing), and that gets us to experimental evidence for aromaticity:

Thanx for the elaborate answer. What I had in mind was the question whether NMR shift predictions can be used to tell me if a certain ring system is aromatic or not, and in case of fused rings, which atoms and which bonds are aromatic and which not. I’m sure the prediction error for 1H NMR shifts is well below 2ppm, and more in the order of 0.2ppm.

But maybe I should be asking, can I use CrystalEye to decide if ring systems are “aromatic”, and in case of two rings fused together (non-spiro), which atoms and bonds are aromatic and which not. Aromaticity is a fuzzy concept, with various definitions. I would be interesting in linking what the expert considers ‘aromatic’ (or SMILES, or the CDK, or …) with what the QM chemistry (via bond lengths or NMR shift predictions) and crystal structures (via bond lengths) has to teach us. The null hypothesis being that the bonds are not delocalized (bond length) and that no ring current is found (NMR shifts, 1H in particular).

Regarding those bond lengths, ‘aromatic’ bonds show a bond length in between that of single and double bonds (e.g. see this random pick). The CrystalEye data does not reflect that really, and only a trimodal histograms shows up. Indeed, the C#C peak is very low, around 1.2A :) Apparently, the triple C#C bond order is underrepresented in nowadays crystallography.

Maybe aromatic C:C bonds are underrepresented too, or can the absence of a peak around 1.40A be explained otherwise? I would at least have expected a shoulder or deviation in peak shape of the peak at 1.37A.

This is what the histogram looks like (for archival reasons):

1. the data set it way too small
2. the proposed proof-of-concept is not novel at all
3. they do not cite me

Well, the latter sounds a bit boohoo, and it is :) (BTW, I do like this paper.)

They propose the work as proof-of-concept, but use a very artificial data set of only 12 crystal structures (benzene and eleven polycyclic aromatic hydrocarbons, like naphtalene, anthracene, phenanthrene, triphenylene, pyrene, perylene, and coronene). While 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 work, I analyzed this problem myself, and published two years ago the paper Method for the computational comparison of crystal structures (DOI:10.1107/S0108768104028344). Apparently, Parkin was not aware of this publication and did not cite it. I should have went to a crystallography conference with a poster, and advertise my work more. In this paper, I analyzed a data set with 48 crystal structures, manually validated by visual inspection, resulting in having to compare 1128! crystal structure pairs. Took me two full weeks behind a Silicon Graphics. Yes, I really understand why they took only 12 structures :)

However, there is more prior art. While my approach was based on a new radial distibution function-based whole crystal structure descriptor, my supervisor (Ron) used the more common powder diffraction pattern and showed in Representing Structural Databases in a Self-Organising Map (DOI:10.1107/S0108768105020331) it to be a good enough descriptor for clustering of thousands of crystal structures using a self-organizing map (SOM).

Last week, my second paper in crystallography appeared: Supervised Self-Organizing Maps in Crystal Property and Structure Prediction (DOI:10.1021/cg060872y). In this paper, we show how supervised SOMs (see DOI:10.1016/j.chemolab.2006.02.003) can be used for supervised classification and even for property prediction. Note that these supervised SOMs are truly supervised, unlike many earlier modifications of the unsupervised SOMs: the training is supervised.

Finally, another advantage of this last work: the code is open source. The code for the unsupervised SOMs is available as R package: kohonen; and for powder diffraction patterns: wccsom. Details can be found in this R News issue. The first package is not actually limited to crystal structures, and can be used for any clustering problem. However, the articles mentioned here make use of simulated diffraction patters, and I am not sure there are open source tools to generate those.

BTW, I would still be interested in teaming up with CrystalEye in one way or another, and couple these data analysis methods to live streams of new crystal structures. Nick, let me know if you are interesting in idea exchange.

Getting back to Parkin’s paper, I do like the work. Hirshfield surfaces are an interesting tool to visualize packing characteristics, and using them to describe a crystal structure sounds like an interesting idea indeed. I just hope that the method properly scales.

Nabuurs discusses in this article the uncertainties that come with NMR derived 3D molecular structures of proteins. These studies do not give factual data on atomic coordinates, but generally give facts about interatomic distances. Solving the 3D geometry is then an optimization problem where the task is to find the 3D geometry that best reproduces the factual interatomic distances.

Now, this optimization has many closeby, i.e. in terms of matching the experimental data, minima, corresponding, possibly, to quite different structures.

This is nicely demonstrated in the article, by comparing the folds of 1Y4O and 1TGQ, as shown in the figure below (CCAL license):

It is interesting to note that 1TGQ got replaced by 2B95 about the same time the article by Nabuurs was published, which shows a 3D model that is homologous with that of 1Y4O, and different from that in the Nabuurs article.

Among these presentations is mine, on comparing crystal structures (PDF ) and deducing cell parameters. But I’m having a poster on QSAR too.

I’ll arrive on saturday afternoon in Goslar, so leave a message at the conference hotel if you want to meet up, and talk about my work, or yours, or the CDK, KDE, JChemPaint, Jmol, kfile_chemical, Kat/Chemistry, BlueObelisk, Eclipse, R, or whatever else… I plan to have a modest german meal and one or two beers in the evening.

BTW, after Belém (Lissabon), Sintra, Boppard, Kinderdijk, Hoorn and Cologne, it’s the 7th UNESCO world heritage site I’m visiting in just 14 months! Can’t we just have conferences in Hawaii and sorts, like they do in other fields?? Oh, wait, we do: EuroQSAR is on a cruise boat.