Corante

About this Author
DBL%20Hendrix%20small.png College chemistry, 1983

Derek Lowe The 2002 Model

Dbl%20new%20portrait%20B%26W.png After 10 years of blogging. . .

Derek Lowe, an Arkansan by birth, got his BA from Hendrix College and his PhD in organic chemistry from Duke before spending time in Germany on a Humboldt Fellowship on his post-doc. He's worked for several major pharmaceutical companies since 1989 on drug discovery projects against schizophrenia, Alzheimer's, diabetes, osteoporosis and other diseases. To contact Derek email him directly: derekb.lowe@gmail.com Twitter: Dereklowe

Chemistry and Drug Data: Drugbank
Emolecules
ChemSpider
Chempedia Lab
Synthetic Pages
Organic Chemistry Portal
PubChem
Not Voodoo
DailyMed
Druglib
Clinicaltrials.gov

Chemistry and Pharma Blogs:
Org Prep Daily
The Haystack
Kilomentor
A New Merck, Reviewed
Liberal Arts Chemistry
Electron Pusher
All Things Metathesis
C&E News Blogs
Chemiotics II
Chemical Space
Noel O'Blog
In Vivo Blog
Terra Sigilatta
BBSRC/Douglas Kell
ChemBark
Realizations in Biostatistics
Chemjobber
Pharmalot
ChemSpider Blog
Pharmagossip
Med-Chemist
Organic Chem - Education & Industry
Pharma Strategy Blog
No Name No Slogan
Practical Fragments
SimBioSys
The Curious Wavefunction
Natural Product Man
Fragment Literature
Chemistry World Blog
Synthetic Nature
Chemistry Blog
Synthesizing Ideas
Business|Bytes|Genes|Molecules
Eye on FDA
Chemical Forums
Depth-First
Symyx Blog
Sceptical Chymist
Lamentations on Chemistry
Computational Organic Chemistry
Mining Drugs
Henry Rzepa


Science Blogs and News:
Bad Science
The Loom
Uncertain Principles
Fierce Biotech
Blogs for Industry
Omics! Omics!
Young Female Scientist
Notional Slurry
Nobel Intent
SciTech Daily
Science Blog
FuturePundit
Aetiology
Gene Expression (I)
Gene Expression (II)
Sciencebase
Pharyngula
Adventures in Ethics and Science
Transterrestrial Musings
Slashdot Science
Cosmic Variance
Biology News Net


Medical Blogs
DB's Medical Rants
Science-Based Medicine
GruntDoc
Respectful Insolence
Diabetes Mine


Economics and Business
Marginal Revolution
The Volokh Conspiracy
Knowledge Problem


Politics / Current Events
Virginia Postrel
Instapundit
Belmont Club
Mickey Kaus


Belles Lettres
Uncouth Reflections
Arts and Letters Daily
In the Pipeline: Don't miss Derek Lowe's excellent commentary on drug discovery and the pharma industry in general at In the Pipeline

In the Pipeline

« A Quick Tour Through Drug Development Reality | Main | Shootmenowicene »

August 15, 2012

More On Reproducing Scientific Results: Organic Chemistry Edition

Email This Entry

Posted by Derek

Yesterday's post on reproducing scientific results got me to thinking about the application of this to organic chemistry. How much of this are we going to see, compared to biology?

Not as much, is my guess. Some of the barriers to reproducibility are too low to bother with, while others are too high. In the "too low" category are many new synthetic method papers. People try these things out, if they look useful at all, and they either work or they don't. Most of the time, you end up finding the limits of the reported method - your substrate failed dismally, but when you look, you realize that you had a basic tertiary amine in your molecule, and none of the examples in the paper have one. Ah-hah.

It's rare that a useful-looking reaction turns out to be completely non-reproducible across multiple structures (although it has happened). Here's a paper from 2000, by one Vincent C. O. Njar, claiming that carbonyl diimidazole reacted with hydroxy groups to give direct N-alkylation of imidazole. Two years later, Walter Fischer from Ciba Specialty Chemicals took this paper apart in detail, showing that it did not work and could not have worked. The products were carbamates instead - not surprising - and the original author should have realized this (as should the referees of the paper).

Then you have total synthesis. And here, the barrier is too high: no one is going to reproduce these things after a certain point. A 48-step synthesis of Shootmenowicene could appear tomorrow, and the odds are overwhelming that no one will ever explore its heights again. There have been total syntheses that have been received with grave doubts (hexacyclinol!), but no one, to the best of my knowledge, has gone back over every step of one of these. The return on the investment of time and money is just too low - which, to be frank, is a sentence that sums up my opinion of a lot of total synthesis work these days.

Where the Reproducibility Initiative could come in handy inside organic chemistry, though, would be for unusual things of wide applicability that are still hard to believe. The famous "NMR chirality" scandal at the University of Bonn in the 1990s would be a good example of this. This was a startling result - that the chirality of organic reactions could be measurably influenced by the handedness of an applied magnetic field - and many people had trouble believing it on physical grounds. They were right, too, because it all turned out to be faked by an individual inside the group, a fact that was only discovered after much effort and embarrassment. Having immediate access to third-party reproducibility testing would have sped things up quite a bit - and perhaps if that access is more widely known, used, and appreciated, we might see fewer bizarre cases like this in general.

Comments (16) + TrackBacks (0) | Category: The Scientific Literature


COMMENTS

1. CanChem on August 15, 2012 3:13 PM writes...

As I was still in high school when both those Synthesis papers came out, forgive me if this would have been unusual at the time, but wouldn't a 13C NMR have sorted out everything in about 5 minutes? That carbamate carbon would be both 10 ppm removed from the rest of the carbons, and would throw off the carbon count...

Permalink to Comment

2. Anon on August 15, 2012 4:11 PM writes...

I'm a Corey guy, reputation of this from the lab. What I can add is that the chem was ridiculously difficult, but we got it to work. Sometimes irreproducibility is due to technical difficulty, not that it doesn't work...

Permalink to Comment

3. Chemist of sorts on August 15, 2012 8:32 PM writes...

Isn't this essentially what Org. Syn. is? Hexacyclinol should have been caught by the refs. Sadly, experience teaches to take much of the literature with a grain of salt. Profound discovereries should be carefully checked anyway... I really don't want to see any more Pd "free" couplings reported until its established beyond any doubt.

Permalink to Comment

4. Chrispy on August 16, 2012 12:07 AM writes...

@2Anon -- Corey was the first lab I thought of when I
saw this. Maybe no one matches your talent but certainly no one matches your claimed yields. There was a joke about the "Corey 1:2 inversion" which made a 29% yield into a 92% one...

Permalink to Comment

5. Samaj-Dewar on August 16, 2012 12:32 AM writes...

The reproducibility may well work product wise but yield claimed by many groups will be very hard to reproduce. This also is a kind of reproducibility issue.

Permalink to Comment

6. Al-chemist on August 16, 2012 2:42 AM writes...

When I was post-doc'ing, I was doing organocatalysis during the years when you could just buy proline, stick it in with a couple of chemicals from Aldrich and stir at rt, then publish in JACS a week later with >90% yield + ee....good times! Owing to the overwhelming speed and simplicity of the procedures, the boss would only believe your results if you gave them to another post-doc who could successfully repeat it. We did the same thing with every paper referee'd from other groups. The result was that most of that work was pretty solid, even when reproducing stuff by the guys with poor reputations ....but then again, some of the Lewis Acid guys were pretty open about limited reproducibility of their results: in their mind, an asymmetric Lewis-acid catalysed reaction was good if it produced the claimed results 1 in every 2 times. But really, by that point it is already becoming quite a headache to attempt reproducing those results. And modern organocatalysis is the same: everybody has diverged into their own little niche area now, and it would be much harder to aufit now.

Permalink to Comment

7. Morten G on August 16, 2012 4:26 AM writes...

Well chemistry does have Org. Syn. ...

If the lab publishing the original article supplied the starting materials and intermediates it wouldn't be too much work?

And to be honest I think it should be up to the journals to check the technical data, e.g. NMRs, mass specs, xrays, and not the referees.

Permalink to Comment

8. newnickname on August 16, 2012 6:12 AM writes...

@2 Anon: I'm a bit confused by your brief post. Are you saying that you DID successfully reproduce Njar's results and obtained N-alkyl imidazoles (from heating the carbamates)?

Hmmm ... Vennerstrom reports (Synthesis, 2006, 2540) some examples that contradict Njar's claims (stable carbamates, not N-alkyls and, @1 CanChem, C13 DOES show the carbonyl carbon) but some that do rearrange to alkyl imidazoles after heating at 170 C.

As an aside, I think Njar is the guy made abiraterone, a recently approved drug for prostate cancer that got some play in Pipeline not too long ago. (September 15, 2011
Terra Slightly Less Incognita)

Permalink to Comment

9. John Li on August 16, 2012 6:44 AM writes...

@1 CanChem

I am not surprise if the C=O signal is too weak to observe. Even the 13C NMR spectrum for CDI shown on the sigma aldrich website, the C=O is just abnormally weak.

However that didn't explain the HRMS and elemental analysis

Permalink to Comment

10. Grignard on August 16, 2012 7:54 AM writes...

Great series of thoughts, Derek! Besides the more obvious traps of non-reproducibility, too many synthetic data are irritating because they are far from being reproducible even in the low gram range. These data are supposed to be "hard", well-defined number, but turn out to be quite "soft": yields, purity, ee, etc. Still, the 99% ee numbers are taken as benchmark by competing groups in the field and by referees. Some groups are aware of the limits of small scale reactions and openly point this out (Hudlicky, Synlett 2012).
My advise for everybody looking for reproducible results is to have a look at Org. Process Res. Dev., the ACS journal reporting kg- and ton-scale recipes. There is no doubt that it works.

Permalink to Comment

11. Anonymous on August 16, 2012 9:13 AM writes...

Grignard/OPR+D: there has been the occasional blemish....!
But it was tremendous comic relief.

Permalink to Comment

12. simpl on August 16, 2012 9:28 AM writes...

On the subject of reproducibility, almost every CAS number I check for existence on the internet brings up sites offering the material for sale. I feel like ordering on occasions, just for fun.

Permalink to Comment

13. Jordan on August 16, 2012 2:05 PM writes...

Reproducibility is the main reason why I (and a few other commenters, it seems) trust Org Syn and Org Proc Res Dev more than the rest of the organic chemistry literature. The chemistry isn't as novel, but it's a lot more reliable.

Permalink to Comment

14. cynical1 on August 16, 2012 2:36 PM writes...

I remember that Njar paper. I remember asking myself, 'How the hell does that work?' But I knew that a colleague of mine was trying to make analogs in that ilk so I pointed out the paper to him. He tried to reproduce the reaction, determined that he had the carbamate and not the aryl imidazole and even contacted the author. Njar replied back the first time and of course refused to admit his failure. Well my friend reproduced one of the examples in the paper and sent him the C13 NMR and LC/MS which showed that he had published crap. The guy never wrote back. I felt really bad for pointing it out to my friend because it wasted his time. If it looks too good to be true, it probably is.

Then I saw the Fisher paper when it came out and pointed it out to my friend. I laughed and told him that Fisher got a paper and a line on his resume out of it and all he got was the wrong product and had to figure out another way to make them. You don't get time in Pharma to publish why other people's published crap is just that. You just ignore it and move on, that is, until someone buys that piece of crap for 700 million dollars and you lose your job to finance it.....but I digress.

Permalink to Comment

15. Anonymous on August 16, 2012 3:47 PM writes...

hm,
so say a guy like buchwald publishes yet another 15-substrate-cat-screening with a limited functional group spectrum using a fancy phosphate, one of his mates checks it and both get a paper and a citation?

the thing with org. syn. is it just gives you the odd lucky shot of a working procedure when your substrate is reasonably close to the published example. they often use more text for conditions and remarks than for the procedure itself. such degree of optimisation and reproducibility is next to impossible for most routine organic chemistry papers.
let alone having reproduce a 1 L distillation or a 1 kg scale recrystallisation from Org. Syn. reports from >2 decades ago with the 100 mg you might have in a late stage of your synthesis...

Permalink to Comment

16. Design Monkey on August 22, 2012 1:30 PM writes...

Vincent C. O. Njar Department of Pharmacology and Experimental Therapeutics, School of Medicine, University of Maryland,

---------

Single author, pharmacology department. Seems to be case of pharmacologist dabbling in chemistry, which he did not understand very well.

Permalink to Comment

POST A COMMENT




Remember Me?



EMAIL THIS ENTRY TO A FRIEND

Email this entry to:

Your email address:

Message (optional):




RELATED ENTRIES
How Not to Do It: NMR Magnets
Allergan Escapes Valeant
Vytorin Actually Works
Fatalities at DuPont
The New York TImes on Drug Discovery
How Are Things at Princeton?
Phage-Derived Catalysts
Our Most Snorted-At Papers This Month. . .