petermr's blog

Totally Synthetic posted his blog stats today (Blogging… ) so I thought I’d post the ones that Jim Downing has set up for us. Apart from anything else I like the graphics.

Days of week
Mon Tue Wed Thu Fri Sat Sun Day Pages Hits Bandwidth Mon 779.50 1365 25.38 MB Tue 4094 7443 161.17 MB Wed 3930 7989.50 149.56 MB Thu 4543 10763 129.98 MB Fri 2473.50 5738.50 92.76 MB

I don’t get this. We do work on Mondays in the UK. It can’t be timezones – e.g. Monday is not really Tuesday. The interesting comparisons are for the next two:

Operating Systems (Top 10) – Full list/Versions – Unknown
Operating Systems Hits Percent Windows 34047 50.6 % Unknown 15733 23.3 % Linux 10519 15.6 % Macintosh 6905 10.2 % Sun Solaris 27 0 % FreeBSD 23 0 % HP Unix 2 0 % GNU 2 0 %

Browsers (Top 10) – Full list/Versions – Unknown
Browsers Grabber Hits Percent Firefox No 24387 36.2 % MS Internet Explorer No 18612 27.6 % Unknown ? 13311 19.7 % Safari No 5226 7.7 % Mozilla No 3650 5.4 % Netscape No 823 1.2 % Opera No 613 0.9 % NetNewsWire No 322 0.4 % Konqueror No 126 0.1 % TelePort Pro Yes 63 0 %   Others   125 0.1 %

Here Totally Synthetic has

Now it is dangerous to compare stats collected on two different systems, but it seems like that (a) there are a greater proportion of chemists reading TotSynth and (b) chemists are more likely to use Windows and MSIE.
If so, that is one of the things we have to take into account when suggesting approaches to software development.

This is a slightly provocative post – I would welcome comments.
When the Nobel prizes were announced this year there was consternation in some of the chemical blogosphere that the chemistry prize had gone to a biologist (see links in my earlier post). But others argued that chemists had lost the plot and that biology is now where the action is. It was important not just to do research in one’s own community but to appeal to a wider audience.
A lot of chemistry (reflected by the chemical blogosphere) is concerned with making compounds. A subset of this is the synthesis of natural products, compounds produced by organisms (fungi, bacteria, plants, marine invertebrates, etc.). These compounds often have biological activities (e.g. to kill other other organisms) and are frequently useful in medicine. Typical examples are taxol (anticancer, from the Pacific yew tree) , ciclosporin (immunosuppressant from a fungus) and vancomycin (antibacterial from bacteria). They are in widespread clinical use and this is reflected by a very large literature. I have used Pubmed to get a base metric – I simply enter the name and see how many references there are and the earliest date:

• taxol (1971) 12505 references
• vancomycin (1955) 12225 references
• ciclosporin (1980) 20505 references

(The numbers are approximate as names can vary – we also find “cyclosporin” – but Pubmed with its MeSH terminology is very good about synonyms).
These compounds are difficult and expensive to extract from natural sources so there is clear value in trying to make them more cheaply by synthetic chemistry. Moreover it may be possible to synthesize similar compounds which are better or much easier to make or both. So not surprisingly the pharmaceutical industry (in which I have worked) and academia are interested in synthesizing complex molecules. And this is reflected, for example, in TotallySynthetic‘s excellent blog where he reviews the syntheses of natural products.
But are the targets chosen because they are useful, or because they are difficult? And do the results of such synthesis find their way into the clinic or use as biological tools? To be fair, it’s important to remember that it takes a long time to develop drugs, and also that chemical synthesis might be a prerequisite step before any other science can be done. Bearing that in mind, here are the Pubchem metrics for the compounds in the last 2 months of Totally Synthetic’s blog (back to 2006-09-01). The table shows:

• the name used in the blog. (Sometimes there are additional letters which I have removed to increase the hits). Bear in mind that some compounds might have synonyms, but Pubmed is pretty good about this.
• The date of the first abstract in Pubmed. This is often the first report of the isolation of the compound and its possible biological activity.
• The number of papers in chemistry journals. These are mostly reporting the synthesis, but some might discuss the elucidation of structure. The latest chemistry paper as abstracted by Totally Synthetic is sometimes not yet included in Pubmed, but in general all the major chemistry journals are reviewed.
• The number of papers in “non-chemistry” journals. This is subjective but these are the ones where biosynthesis, biological activity, medicinal properties, etc. are likely to be reported.

Out of 22 compounds a few have a significant biological literature. Many, even when reported some time back have none. So it is difficult to argue that they are synthesised because they are interesting.
There are other reasons than interesting biology for doing this type of synthetic chemistry. It can lead to new synthetic methodology which could be of use to the chemical industry in general. And it may be good training for the chemistry-specific part of a PhD. So I welcome comments – they will all be posted.

In many detective novels the murderer makes a fatal mistake. Sometimes they don’t realise this, sometimes they try to cover their tracks. On the Internet you can’t cover your tracks –

“The Moving Finger writes: and, having writ,

Moves on: nor all thy Piety nor Wit

Shall lure it back to cancel half a Line,

Nor all thy Tears wash out a Word of it.”

So there is an inadvertent half a line in the history (actually less) that I cannot wash out and is a glaring pointer to the solution. I’ve only just spotted it.
However I promised that if people wrote comments I might publish more clues. So here is the latest comment:

1. Propter Doc Says:
   November 10th, 2006 at 5:12 pm eWell I spent a good bit of last night trying to figure things out, but never got anywhere close enough to post a comment. I would say that there was very little logic in my approach which ranged from guessing ‘orange stuff we’d encountered as undergrads’ and looking up wikipedia, to searching web of science for your previous papers to see if there were any clues there.
2. I replied
   November 11th, 2006 at 12:34 am e(1)Thanks very much for posting. Although everything I reported in the original post was correct, the molecule is a bit of a chameleon and dark red is not its commonest colour which is yellow. I’ve probably acted like too many murderers, priding themselves on their undetectability, while gradually giving away more information. There is now enough information on these pages to solve it, given all that has been written and to know that the result is almost certainly correct

So, to confirm this statement here is a picture of a crystal of the mystery molecule (not real size). I am not breaking copyright by posting this (though I can’t say why as that would give more clues). It’s much more beautiful than this suggests.

I can’t think of any more clues at present. But if I get more mail I might – or I might not.

It’s three days since the mystery molecule was announced and although I’ve had local interest in Cambridge there’s only been one external post. You must know that in police work it is critical to make rapid progress in the first 48 hours. After that the trail gets colder. So if you had fed me with comments, there might have been some more clues.
I have confirmed that Peter’s approach works. It actually mirrors much police work – careful systematic collection of large amounts of small pieces of data. If this were a matter of national interest or there were a finiancial reward it would have been solved by now.
Remember that your posts can be anonymous. And not to give the mystery away.
But perhaps it’s simply that it’s not worth blogging chemistry – I have an interesting Open Access item that I must post later today and will certainly get some feedback.
P.

There has certainly been some speculation in our real-life community about the molecule! I am assuming that there are some other chemists out there who are speculating. Please give us some feedback – if you do maybe there will be some more clues – or some more bodies…
Two public feedbacks from my colleagues:
Peter Corbett blogs

Jim Downing is stuck on a mystery molecule, and wonders whether the answer can be found via search. Well, I doubt that conventional chemical informatics search will do him. Google Suggest, on the other hand, may be his friend. It seems that Peter Murray-Rust is being coy about a certain critical piece of information, something that’s normally unimportant that may be considered to be a dead giveaway. Well, it’s a giveaway because of linguistics. You see, there’s a couple of collocations that describe the molecule in question, and if you can guess one half of the collocation – or systematically search through the available values (there’s a finite set, and not too large) using publicly available tools trained on the entire web as a corpus – then the other half will leap right out at you.
I think there’s a broader point to make here, but that will have to wait for another time. There’s an idea I’d like to talk to real people about before I blog it.

Remember that Peter knows what the answer is. Although I am holding stuff back (what detective writer doesn’t?) I am not sure what particular bit he’s referring to. (I admit only to the space group and the cell dimensions – that is all I knew in 1970 when I first encountered the mystery molecule). But Peter is right that “the answer is out there” – Judith MR disovered a splendid site today which gives the complete story.
Jim Downing blogs:

I’m whiling away a few minutes trying to work out Peter’s mystery molecule. IANAChemist, and although I’ve been working with them for 10 months now, I fear that very little has rubbed off. However, Peter reckons that you can’t work it out chemically, so I figure that not being a chemist shouldn’t be too much of a handicap. Like betting on the novices at Newmarket.
I leapt off to Pubchem, and used the molecular weight limits. Restricted to 240-260 (the syntax is 240:260[mw]). Nearly 14,000 compounds. Hmmm. Narrowing it down, 245-255 gives under 7000. Still needle in a haystack.
Need to assess what we know, and make some inferences. Hopefully the lazy web will mean some kindly chemist can fix my broken logic.
OK, so what do we know? It seems to be fairly commonly observed, so I’m guessing its structure was discovered a while ago. Well, you can’t limit in PubChem on a range of CIDs, but you can limit to NCI as the source of the data. This didn’t reduce the hit count.
We also know that Peter was trying to do co-ordination chemistry at the time. I didn’t know what that co-ordination chemistry is (I told you, I’m really not a chemist), but wikipedia soon sorted me out. So there are probably either metal ions or organic ligands. My A level chemistry and the variety of colours of the crystals suggest that the metal ions are in the molecule. So I added a limit requiring a heavy atom into pubchem (I know this doesn’t necessarily mean metal ions). Only 14 Hits, whoppee! Probably the wrong 14, but hey.
So my best guess is molecular iodine, which Peter assures me is wrong.
Where is my logic broken? Is it possible to find this molecule by search?

This is a brilliant approach and worthy of the Web. Unfortunately not all detective mysteries are solved by logic. So, for example, most chemists will know immediately that they have molecular iodine (it’s dark metallic grey and purple in solution). So I suspect that Jim cannot know that he has got the right answer, whereas anyone who has come across this phenomenon will. Whether a chemist can “work it out” or search for it from the clues given I don’t know. Peter thinks yes.