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: Twitter: Dereklowe

Chemistry and Drug Data: Drugbank
Chempedia Lab
Synthetic Pages
Organic Chemistry Portal
Not Voodoo

Chemistry and Pharma Blogs:
Org Prep Daily
The Haystack
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
Realizations in Biostatistics
ChemSpider Blog
Organic Chem - Education & Industry
Pharma Strategy Blog
No Name No Slogan
Practical Fragments
The Curious Wavefunction
Natural Product Man
Fragment Literature
Chemistry World Blog
Synthetic Nature
Chemistry Blog
Synthesizing Ideas
Eye on FDA
Chemical Forums
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
Gene Expression (I)
Gene Expression (II)
Adventures in Ethics and Science
Transterrestrial Musings
Slashdot Science
Cosmic Variance
Biology News Net

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

Economics and Business
Marginal Revolution
The Volokh Conspiracy
Knowledge Problem

Politics / Current Events
Virginia Postrel
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

« Not the Usual Morning Around Here | Main | Cancer: Back to N-of-One »

April 22, 2013

Real Reactions, From Real Lab Notebooks

Email This Entry

Posted by Derek

Over at NextMove software, they have an analysis of what kinds of reactions are being run most often inside a large drug company. Using the company's electronic notebook database and their own software, they can get a real-world picture of what people spend their time on at the bench.

The number one reaction is Buchwald-Hartwig amination. And that seems reasonable to me; I sure see a lot of those being run myself. The number two reaction is reduction of nitro groups to amines, which surprises me a bit. There certainly are quite a few of those - the fellow just down the bench from me was cursing at one just the other day - but I wouldn't have pegged it as number two overall. Number three was the good old Williamson ether synthesis, and only then do we get to the reaction that I would have thought would beat out either of these, N-acylation. After that comes sulfonamide formation, and that one is also a bit of a surprise. Not that there aren't a lot of sulfonamides around, far from it, but I was under the impression that a lot of organizations gave the the semi-official fish-eye, due to higher-than-average rates of trouble (PK and so on) down the line.

My first thought was that there might have been some big and/or recent projects that skewed the numbers around a bit. These sorts of data sets are always going to be lumpy, in the same way that compound collections tend to be (and for the same reasons). The majority of compounds (and reactions) pile up when a great big series of active compounds comes along with Structure X made via Reaction Scheme Y. But that, in a way, is the point: different organizations might have a slightly different rank-ordering, but it seems a safe bet that the same eight or ten reactions would always make up most of the list. (My candidate for number 6, the next one down on the above list: Suzuki coupling).

There's also a pie chart of the general reaction types that are run most often. The biggest category is heteroatom alkylation and arylation, followed by acylation in general. By the time you've covered those two, you've got half the reactions in the database. Next up is C-C bond formations (there are those Suzukis, I'll bet) and reductions. (Interestingly. oxidations are much further down the list). That same trend was noted in an earlier analysis of this sort, and nitro-to-amine reactions were thought to be the main reason for it, as seems to be the case here. There's at least one more study of this sort that I'm aware of, and it came to similar conclusions.

One of the things that might occur to an academic chemist looking over these data is that none of these are exactly the most exciting reactions in the world. That's true, and that's the point. We don't want exciting chemistry, because "exciting" means that it has a significant chance of not working. Our reactions are dull as the proverbial ditchwater (and often about the same color), because the excitement of not knowing whether something is going to pan out or not is deferred a bit down the line. Just getting the primary assay data back on the compounds you just made is often an exercise in finger-crossing. Then waiting to see if your lead compound made it through two-week tox, now that's exciting. Or the first bit of Phase I PK data, when the drug candidate goes into a person's mouth for the first time. Or, even more, the initial Phase II numbers, when you find out if it might actually do something for somebody's who's sick. Now those have all the excitement that you could want, and often quite a bit more. With that sort of unavoidable background, the chemistry needs to be as steady and reliable as it can get.

Comments (21) + TrackBacks (0) | Category: Life in the Drug Labs


1. JK on April 22, 2013 8:38 AM writes...

A non-chemist here. According to Wikipedia Buchwald-Hartwig amination, now the number one reaction, was discovered (or perhaps discovered by the mainstream) in 1994. How much of the chemistry you use was discovered in, say, the last 20 years?

Permalink to Comment

2. Jose on April 22, 2013 8:55 AM writes...

Huh, how much you want to wager there's a reaallly nice correlation between the rise of dead-easy Pd couplings and pharma going to hell in a handbasket?

Permalink to Comment

3. rtw on April 22, 2013 9:03 AM writes...

I have said for a number of years that the best place to look for likely to be successful reaction conditions on you project is to search the E-Notebook and see what your colleague's have done. They are likely working on very similar systems and chemistry, though likely with different heterocyclic cores. Like you have said exciting isn't what Medicinal Chemists are looking for. This is a valuable resource that is under used in many large organizations in my opinion. Collective wisdom is very good knowledge, and can't always be easily reconstructed from Scifinder searches. The added advantage is you can walk down the hall or make a phone call to your colleague and talk to them directly if you have any questions!

Permalink to Comment

4. MTK on April 22, 2013 10:10 AM writes...


There's also a very good correlation between Pharma going to hell in a handbasket and the distance between the sun and Pluto which came to perhelion in 1989.

Permalink to Comment

5. David Formerly Known as a Chemist on April 22, 2013 10:25 AM writes...

@2, pharma's gone to hell in a handbasket because a) we're focusing on difficult biological targets for disease processes we don't understand very well, 2) the regulatory agencies have set much higher hurdles for success, and 3) payer systems demand much greater evidence of benefit before they agree to reimburse a new drug. Mismanagement and continual reorganizations haven't helped either.

Blaming the woes of the industry on wider availability of an extremely useful synthetic method is one of the funniest things I've read in a while. Thanks for bringing a smile to my face.

Permalink to Comment

6. tocket on April 22, 2013 11:22 AM writes...

It would be interesting to know the average yield for a certain reaction, as extracted from e-notebooks. That would likely give a more realistic number compared to what you read in the literature.

Permalink to Comment

7. Anonymous on April 22, 2013 11:28 AM writes...

Why the surprise - reducing nitros to anilines is probably the second best way (after buying) of making a reagent for the BH or acylation/sulfonylation

Permalink to Comment

8. Yancey Ward on April 22, 2013 12:54 PM writes...

The Buchwald/Hartwig brought a smile to my face. It was the subject of the first paper I published as an employed pharmaceutical chemist in 1996- the solid-phase version, of course- during the combi-chem craze.

Permalink to Comment

9. Curious Wavefunction on April 22, 2013 1:07 PM writes...

I think #2 has a point. Even if the wide availability of cross-coupling reactions is not a major factor for pharma going to hell, it has encouraged chemists to do the easy reactions and generate more of the same.

In fact in a very interesting 2011 J. Med. Chem. paper (click on my handle), Walters et al. made the same point and encouraged chemists to move away from cross-coupling reactions.

"Reduced reliance on “easy” chemistry: While Pd-mediated
sp2-sp2 couplings and amide bond-forming reactions
have their place, we believe that a greater emphasis on
the art of synthesis in medicinal chemistry would dramatically improve the physical properties of our molecules."

Permalink to Comment

10. bad wolf on April 22, 2013 1:24 PM writes...

Well, how much you want to wager there's a nice correlation between the focus on a Total Synthesis background from a fancy group for hiring and dead-easy Pd couplings plus amide synthesis ("dull as the proverbial ditchwater") being the actual job they're asked to do?

Permalink to Comment

11. Anonymous on April 22, 2013 2:31 PM writes...

@10. One would think there would be much less of the "never fail dull as ditchwater" chemistry with "fancy group background" employees. Again one would guess they have a bigger repertoire of reactions and would have been trained to work on harder problems. Hence, the reasons they were hired in the first place. The drawback is their tendency to wish to pluck out their eyes after they have been forced to run yet another sulfonamide coupling.

Permalink to Comment

12. David Formerly Known as a Chemist on April 22, 2013 3:57 PM writes...

@11. It doesn't matter what you've been trained to work, what matters are the metrics by which you're evaluated when you land in the industrial lab. So what if you spent 5+ years working out a total synthesis of please-blow-my-brains-out-nowamine, and creating a new reaction methodology or two in the process? When management measures your productivity by the number of compounds you register per quarter, or how many reactions you log into your electronic lab notebook per month, or how many discreet operations (reaction setup, workup, chromatography, etc) you conduct in a week...

What, sounds horrible? I've seen these metrics proposed and used to varying extent when I worked in industry. So the hot shot new hire from name-your-star-chemist's lab quickly learns he doesn't win awards for spending two months working out how to synthesize that sexy tricyclic heteroaromatic with 3 chiral centers. He finds out he hits his annual review goals by preparing a crapload of sulfonamides, amides, and other parallel-izable reaction products. Watching managers treat medicinal chemists like widget manufacturers was one of the saddest things to witness, really sucks the soul out of you.

Permalink to Comment

13. Jose on April 22, 2013 6:12 PM writes...

Here's the scenario I was picturing:
1) MBA-driven productivity metrics get instituted
2) Chemists are rightly unsure about job security- how to rapidly and readily explore some SAR space to get a higher metric?
3) Cranks out 5-10 'interesting' compounds via Pd couplings
4) One looks promising, so a small side project begins
5) More Pd couplings
6) Conventional wisdom about the project SAR evolves based on the flat-ish structures from couplings
7) Project's entire exploration of SAR space gets re-directed away from more complex heterocycles or scaffolds
8) Project tanks from tox or hERG or in Phase i/II

Permalink to Comment

14. Add N to X on April 22, 2013 10:04 PM writes...

The tiny amount of oxidations came as a surprise: I guess we're content to leave those to the body?

Permalink to Comment

15. London_Chemist on April 23, 2013 3:35 AM writes...

@14. Small number of oxidations because:

Disposing of Cr waste (or any other heavier metal like Ru) is a right pain in a Pharma setting (lots of paperwork) so that leaves only a few methods, some of which have really long safety assessments (more restrictions). Its just easier to find an alternative route.

Permalink to Comment

16. Morten G on April 23, 2013 5:23 AM writes...

Why would this be boring to academics? The people making synthesis suggestion software get an idea of what reactions they have to play with - the people making new reactions can estimate the approximate chemical space accessible with this chemistry and attack problems that are actually interesting to industry.

Permalink to Comment

17. Andy on April 23, 2013 10:32 AM writes...

Perhaps there are also so many Buchwald Hartwigs because they are reliable, but slightly fickle and substrate dependent with many different conditions? So you may do a very quick small screen of various catalysts on a very small scale - hence 5 times the number of B-H reactions. Or your first doesn't work, but you know it should - so then you do the screen. Or is there now general agreement on the best go-to works-for-all method?

Permalink to Comment

18. BioBritSD on April 23, 2013 2:01 PM writes...

@Andy certainly has a good point. Adding an extra dimension to the visualization to show/filter by yield (and perhaps scale) would be valuable.

Great poster and great work. There is fantastic data in those ELNs just waiting to be mined. ELN search tools have hitherto been focussed on the "give me an experiment that contains...." type question. But the capability to measure and analyse higher level trend, or chemical space, or....? properties is really interesting (to me).

Permalink to Comment

19. Mycroft W on April 23, 2013 2:02 PM writes...

First read that one line as:

"Then waiting to see if your Pb compound made it through two-week tox, now that's exciting."

I would imagine so.

Permalink to Comment

20. Noel O'Boyle on April 23, 2013 3:55 PM writes...

Hi, Noel O'Boyle from NextMove Software here. Thanks for the write-up Derek, as well as the interesting comments everyone.

The work presented was just an initial analysis, so we didn't want to go beyond the top 5 (we will at a later date), but Derek's prediction for the number 6 reaction is spot-on: Suzuki coupling.

@BioBritSD (18): Couldn't agree more. The whole idea behind the software is to liberate the ELN information so that companies can do whatever sorts of analyses they can think of. We are just scratching the surface with reaction naming.

Permalink to Comment

21. Add N to X on April 24, 2013 12:09 AM writes...

@15. Small number of oxidations because:

There's hope! A just published paper from GSK on their Green Reagent Guides: link includes several on Oxidation, tackling the very issues you mention:

Permalink to Comment


Remember Me?


Email this entry to:

Your email address:

Message (optional):

The Last Post
The GSK Layoffs Continue, By Proxy
The Move is Nigh
Another Alzheimer's IPO
Cutbacks at C&E News
Sanofi Pays to Get Back Into Oncology
An Irresponsible Statement About Curing Cancer
Oliver Sacks on Turning Back to Chemistry