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

« Receptors: Can't Live With 'Em, Can't Understand 'Em | Main | Backtracking, Necessary and Unnecessary »

July 18, 2008

Lowe's Law of Diurnal Distribution

Email This Entry

Posted by Derek

Here’s an appropriate topic for a Friday, although at first many of you may think I’ve lost my mind. What would happen if you combed the full text of the experimental sections of the chemistry journals, looking for how long people ran their reactions?

I’m pretty sure that I know what you’d see: there would be a lot of scatter in the short time periods, with some peaks at the various half-hour and hour marks just for convenience. But as you went out into the multiple-hour procedures, I feel sure that you’d see pronounced spikes in the data at around sixteen to twenty hours and again at around 72 hours.

Some readers have doubtless started nodding their heads, having done the math. Those times correspond to "overnight" and "over the weekend", and I'm willing to bet that they're over-represented (and how) in the data set. I'll go on to predict scarce examples in, say, the 14-hour or 38-hour ranges - there's not much way to run a reaction for those intervals and not be in the lab too early in the morning or too late at night.

A second-order prediction is that when such reactions are found, that their origins will skew heavily toward academia rather than industry. And I'm also willing to bet that patent procedures will tend to follow the working-day timelines more than the general literature, for the same reasons. My last higher-order prediction is that the reaction times would not, in fact, obey Benford's Law, as many other data sets of this kind do.

As far as I know, no one's ever done this sort of analysis, but I suppose it would be possible, especially for someone at Chemical Abstracts or at one of the scientific publishers. If someone wants to try it, please let me know what comes out. And if the results follow my predictions, please feel free to refer to the title of this post or something similar. I won't object.

Comments (31) + TrackBacks (0) | Category: Academia (vs. Industry) | Life in the Drug Labs | The Scientific Literature


COMMENTS

1. Don B. on July 18, 2008 7:50 AM writes...

Eighteen hours was always my standard unless the reactants had a chance to "get exciting".

Permalink to Comment

2. Anonymous on July 18, 2008 8:09 AM writes...

I'll bet 1 hour and 8 hours are overrepresented in the short times, and overnight (not even a numerical time) is the overall winner.

What do I win?

Permalink to Comment

3. skyywise on July 18, 2008 8:37 AM writes...

Working in industry, we had an automated platform that ran up to 1000 reactions per tube, but required at least hourly feeding of samples / tubes. With sample prep time, the experiments would run a good 14 hours... we were at work far to early or late every day.

Permalink to Comment

4. philip on July 18, 2008 8:46 AM writes...

This is an observation that most lab supervisors make eventually. Your humble assistant sets up a reaction that you're really interested in and you ask how it's going (what does the TLC or LC/MS look like?). They say, "I haven't looked yet, it's supposed to go for 16 hours." At this point you have to explain that Lowe's Law of Diurnal Distribution is at work here and that they should take an aliquot now. Often the reaction is done and you look like a wise chemistry sage, when in reality, you're just trying to get the work done.

Permalink to Comment

5. Still Scared of Dinosaurs on July 18, 2008 9:00 AM writes...

You'll notice the same thing when you scale up to clinical trials. Drugs are dosed 1xdaily if possible, then if necessary 2x, then 3x or weekly, monthly, yearly. Assessments are done daily, weekly, monthly, quarterly, yearly. When this absolutely won't work, like for most PK studies, you keep the subjects on the ward until the sampling times span a day. Of course, to do the trials you have to get people in the clinic so addressing human factors in scheduling make sense, but it's rare that you will hear a lot of discussion about what types of effects these apparently arbitrary decisions will have.

What I don't get in the chemistry world is why you aren't expected to sample the reaction throughout its course to measure the kinetics.

Permalink to Comment

6. eugene on July 18, 2008 9:11 AM writes...

20 hours would be a regular one for me a while ago. Set it up in the early afternoon or right before lunch, check it before going back home and find out it wasn't complete, then quench it in the morning.

These days I just say overnight. Most of my stuff is in very small amounts and is too valuable and air sensitive to waste the risk of stopping the reactions and taking an NMR (hard to get away with an aliquot when you only start with 5 mg.) after evaporating it. Even if the reaction is done after one hour (and some of them are when I do them in an NMR tube), I've learned that the risk is there that I'll waste a lot of time and sanity if I don't leave it overnight. Or at least for 8 hours.

Permalink to Comment

7. DrSnowboard on July 18, 2008 9:11 AM writes...

"What I don't get in the chemistry world is why you aren't expected to sample the reaction throughout its course to measure the kinetics."
Because most of the time, for simple medchem reactions, it's an endpoint assay ie has it finished yet? We leave the fancy stuff to the process chemists unless the product is unstable or equilibrates to another product. And I tend to avoid those reactions if I can... There was a lovely paper by Boehringer on the RCM to make BILN2061 a while back, really took the reaction apart.

Permalink to Comment

8. eugene on July 18, 2008 9:15 AM writes...

"What I don't get in the chemistry world is why you aren't expected to sample the reaction throughout its course to measure the kinetics."

Coming from someone who does measure kinetics from time to time because I have very good reason to since it's part of my job to figure out mechanisms, if I was a boss in a total synthesis lab and my student insisted on measuring kinetics of every Suzuki coupling and Lewis Acid catalyzed Michael addition, they'd get a very severe reprimand. If I was their boss in a pharma business, and they did all that and built nice kinetic graphs, they'd run the danger of being fired. Unless you're in process development, you have no business wasting company time like that.

Permalink to Comment

9. Jose on July 18, 2008 9:47 AM writes...

I have a sliding scale myself. For highly valuable or labile substrates, or if the reagents are less than gentle, I'll set up early in the morning, and check 30 min, 2hrs, ~6 hrs, and then overnight. Otherwise, set up 2-3 reactions late afternoon, and let 'em rip overnight. So, yes, 16 hrs wins by a landslide.

Permalink to Comment

10. Spike on July 18, 2008 9:51 AM writes...

You also see the same phenomenon in PK studies. A standard design for oral administration is 0, 1, 2, 3, 5, 8 and 24 hours. The latter part of the curve is purely based on convenience.

As pointed out by Still Scared of Dinosaurs you see some strange things in terms of dose administration. One thing that they didn't mention was the "twice a week dosing" as is used for Velcade. The doses are administered on a 3 day/4 day cycle which is based on convenience rather than any hard core scientific principles.

When you go to b.i.d. dosing, you start to see a diversity of approaches. Some people will use a 12:12 schedule and others will use an 8:16 schedule. In the latter case the animals (dogs in this case) were dosed at 8 am and 4 pm so that the animal techs could go hope at a reasonable time. No problem with that. The dogs were allowed access to their food from 10 am to 2 pm as usual. No problem with that. Based on the doses used, we were anticipating that the exposure after the second dose would be greater than after the first dose. Naturally, the converse was true. We had run slap bang into a negative food effect (for the first dose the dogs were in the fasted state and for the second dose the dogs were in the fed state). Going to a 12:12 schedule increased the exposure in the study. That was one of those Homer Simpson moments (D'oh!).

Permalink to Comment

11. Russ on July 18, 2008 10:57 AM writes...

Ignorance of reaction kinetics and mechanism in most medchem organizations has been the norm for some time now; most medchemists view ovenight reactions as making better use of their time. Interest in chemistry rather than headlong pursuit of the next analog leads out of discovery and into chemical development.

Permalink to Comment

12. CMC guy on July 18, 2008 11:32 AM writes...

Process/Manufacturing has similar if slightly different "time sequence standards" because if often takes longer to set-up and initiate runs then have to work around operator break schedules and shift changes (avoiding anything critical on 3rd shift). Plant time expensive so usually do model or in-process monitor for shortest requirement. Engineers are especially sensitive to importance and economic impacts.

#8 eugene is correct that kinetics worthwhile in process development although in most cases only a gross level knowledge satisfactory (sometimes vital if competing side reactions possible). Understanding thermodynamics needed too so in spite of distaste by (most?) syn people for P-chem it can be useful after all.

In grad school 96h rxn meant a holiday weekend you didn't work.

Permalink to Comment

13. processchemist on July 18, 2008 12:41 PM writes...

eugene:
only applications in process chemistry the NEED kinetics data are reactions in continous reactors. A gross knowledge is all you need for typical batch processing (most of the times).

CMC:
a little more of p-chem culture in discovery guys would not be so bad. I'd like to hear no more statements such "250 mmHg of residual pressure of MeI (off the vent of a vacuum pump) are not so much..."

Permalink to Comment

14. Still Scared of Dinosaurs on July 18, 2008 12:55 PM writes...

BTW I wasn't being critical, just expressing ignorance as to why kinetics isn't an issue.

Also, since it's very common to have measurements on clinical trials scheduled weekly +/- days...outside that window and it goes down as a protocol violation. We commonly suggest that the first dose to a patient be delivered on a Wednesday so that all 5 weekdays of subsequent weeks are within protocol. However, when we get to the last visit for the last patient we want to get everything closed out as quickly as possible. Therefore if anyone were to do the analysis they'd probably see first date of clinical activity weighted towards Wed and last date weighted towards Monday.

And the first date of clinical activity in Europe will rarely be in August or December.

Permalink to Comment

15. Anonymous BMS Researcher on July 18, 2008 1:45 PM writes...


Back in grad school, a fellow student was doing an experiment that required making some measurements every hour for N hours, where N exceeded 15, the sort of thing we'd do at my shop with robotics now, but (1) this was back in the 1980s and (2) if such robots had been available the dollars would not have been. So he camped out in the lab with a sleeping bag and alarm clock. Well, about 4AM a cleaning woman saw him -- and his general appearance was even scruffier than the grad student average -- so she thought he was a homeless man and called the campus cops. The upshot was a 5AM phone call to a sleepy Professor to verify that this guy really was running an experiment.

Permalink to Comment

16. chemdaddy on July 18, 2008 2:29 PM writes...

Get a TLC as soon as everything is in solution. Check again at 30-60 minutes. Crudely estimate a reaction half-life if you can. 7 or 8 half-lives means 99% done in a perfect world. Most likely, your reaction doesn't need to run any longer than this.

Reporting 72 hours when it took 4 is okay so long as the scientist makes the effort to indicate that the three days of spin time had more to do with convenience than with kinetics.

Sadly, many folks will trust the average experimental section that promises a 90% yield after 48hrs when in reality, that same 90% might have been ready after an hour and then pushed to the next step that same day.

Casual data reporting trends are on the rise in my opinion and I don't think this is a good thing for scientific integrity overall.

Permalink to Comment

17. CMC guy on July 18, 2008 2:53 PM writes...

processchemist excellent point about continuous reaction systems.

I don't begrudge discovery (chemists) slipping on stuff like you mention (even though seems rather basic) as they have own headaches/focus as illustrated in the previous couple of posts from Derek.

Permalink to Comment

18. Jim on July 18, 2008 4:02 PM writes...

I think some of us forget what an experimental section is for. It is a record of what was actually DONE. Not when the reaction was finished, or when the optimal product ratios were obtained but a accurate portrayal of what was done by the operator.

I don't really care whether the reaction is finished earlier than it says - I can find that out for myself, or do we now trust blindly in the literature (now there's a post or even a blog all by itself)? What I want is for the reaction to do what it says.

This said, it would be nice if in the text of the paper they justified why they used the times/conditions they did.

Permalink to Comment

19. OrgChemPostdoc on July 18, 2008 4:18 PM writes...

Probably microwave-technology will change the picture a little bit. Overnight will turn into
10 min. At least on small scale. Process is a different world.

Permalink to Comment

20. zts on July 18, 2008 9:44 PM writes...

In my notebook, I always record how often I checked the LC-MS or TLC. If I know that I checked the 20 h reaction at 10 min, 1 h, 4 h, and 20 h, that tells me a lot more than if I only checked it at 20 h. And I usually indicate how fast it seems to be going--i.e., was it done after 4 h but I just didn't have time to work it up that day; was there only 10% conversion after 20 h but that was good enough--useful things to know if the reaction is being repeated. Of course, since that sort of thing doesn't go into the actual publication experimentals, it is really only helpful for me and anyone that pulls up my actual notebook pages.

Permalink to Comment

21. vasili on July 19, 2008 3:03 AM writes...

If the experimental part tells you to stir overnite the reaction is probably finished in 2 h, but at least it shows you that the product has not decompose in that time, and if you repeat the same conditions you'll ge the stuff.

I never trust any experimental at first sight. Always check (if I can) after coffee break, lunch break, etc and if the reaction is complete I do the work up after coming.

It's very useful to check reactions soon in order to optimize conditions. If it goes badly in 1 hour at 100C (TLC or LC-MS ) why not run another expeiment at 50C before leaving home.

Permalink to Comment

22. Norepi on July 19, 2008 9:08 AM writes...

See, for me, most of my reactions that take 16 h or whatever *really do* take 16 h, TLC more oft than not tells me the reaction has begun within 15 min, but just takes a whacking long time to go all the way. So I usually just set these things up at 6PM and go home for the day.

I had a really awful cyclodehydration sort of thing I was running several months ago that took, I kid not, 36 h to go to completion.

But sometimes you do have to wonder, e.g "Does it really take three whole days to cleave that methoxy group?

Permalink to Comment

23. sjb on July 19, 2008 10:46 AM writes...

"So I usually just set these things up at 6PM and go home for the day" (Norepi)

Is there not a culture of making sure a reaction is stable before disappearing for the night / weekend / .... ? So typically the latest I'd stick a reaction on (in industry) is 5pm or so, unless it's something I had done several times before starting earlier in the day, and under relatively mild conditions (like stirring at rt, under not necessarily inert conditions)..?

S

Permalink to Comment

24. zts on July 19, 2008 1:49 PM writes...

sjb wrote "Is there not a culture of making sure a reaction is stable before disappearing for the night / weekend / .... ?" For a lot of reactions you have a pretty good intuition about whether or not it will be stable overnight, based either on previous experience with the reaction type, or with the stability of other analogs, or with literature procedures. But, speaking for myself, I do tend to err on the side of checking the LC-MS before I go home, so if it goes to crap overnight I have some basis for determining whether it might have been OK for shorter time, or whether I should just try something else.

Permalink to Comment

25. MTK on July 21, 2008 9:11 AM writes...

The mechanics (when, how, and how many) of setting up a reaction or reactions really depends, for me, on what I'm trying to accomplish at that time.

If I'm on a fishing expedition for a good starting point for a given transformation, for example, I'm going to set up as many different types of reactions as I can and check them at some time point. It may be overnight, it may be five hours, it may be 30 minutes. For me it doesn't really matter. I'm just looking for product and trying to screen as many conditions as possible, so yes, I'll set up six reaction near the end of the day and let them go.

After ID'ing one or two promising conditions, then I'll start looking at each of the reactions carefully.

Permalink to Comment

26. RTW on July 21, 2008 10:58 AM writes...

All I can say is this is very amusing. Most of the time I would set up reactions to go over night. In the majority of cases they where done and I could spend the next morning working them up, to set up additional ones that evening before I left. Yes - There are plenty that get done in an hour or two, and sometimes I would set them up in the morning, work them up quickly and take the materials on. Additionally I would often times run TLC at intervals to determine when a reaction was complete when I had molecules that would degrade if left for prolonged periods under reaction conditions. That was more of a problem than how soon something was done. Could have several reactions be wasted because they set overnight, and a not so stable protecting group comes off.....

Experience and expedience on the Medchem Synthesis lab really determines the times. Meeting schedules and a myriad of other distractions can also be a problem in addition to ill equipped poorly trained BS/MS scientists doing the work. Very few coming out of training programs now a days can run a TLC for example, so following a reaction is too time consuming for them running a 30 min HPLC/MS periodically.... Which in my experence seemed to be the prefered method by these younguns... BTW I am a BS scientist w 20+ years of bench experence. Too expensive to most organizations that don't value experence but want temporary enthusiasm instead, that they can just toss off a few years later and again hire cheap hands. Glad I left that rat race though I do miss the research.

Permalink to Comment

27. Norepi on July 21, 2008 5:34 PM writes...

sjb wrote "Is there not a culture of making sure a reaction is stable before disappearing for the night / weekend / .... ?"

When I'm simply displacing a benzylic halide with an amine to make some sort of analogue, and I've done it 30 times before, I'm pretty sure it'll work. When in doubt, I'll check a TLC and then scram for the day.

Permalink to Comment

28. zts on July 21, 2008 6:13 PM writes...

In response to RTW: "a problem in addition to ill equipped poorly trained BS/MS scientists doing the work. Very few coming out of training programs now a days can run a TLC for example, so following a reaction is too time consuming for them running a 30 min HPLC/MS periodically."

I'm not sure whether I should address your post or not, since it sounds like you are trolling, or maybe just poorly informed. Anyway, I have found that training varies significantly at all levels, and I have seen many Ph.D.s with questionable skills. But I don't really want to argue about that. Why should you lament the use of LC-MS, out of all the more valid things to be grumpy about? I don't know where you get 30 min from--all the methods I have used have taken 5-10 min, which is almost the same amount of time that it takes to run a TLC, and you get a much more quantitative idea of reaction progress, and, importantly, it usually tells you whether your reaction is working or not, much more reliably than TLC, which saves an immense amount of time since you do not have to work up a reaction that did not work. In those cases where LC-MS is not effective, I don't know anyone who doesn't know how to run a TLC. Most people are still trained on TLC, because most graduate programs do not routinely use LC-MS (AFAIK). As you say, expedience is important in medchem labs, so I would say that people *shouldn't* use TLC.

Permalink to Comment

29. DrSnowboard on July 22, 2008 3:26 AM writes...

Only problem with TLC is that it shows you all the impurities that your 3min LC-MS runtime conveniently obscured...'More quantitative' is a good description.

Permalink to Comment

30. T on July 22, 2008 8:06 AM writes...

A fellow co-worker told me once that Americans are more apt to write reaction times in journals in lengthier times or 'overnight', whereas Japanese chemists are more apt to report the reaction time down to the hour and even minute. Is there any truth to this?

Permalink to Comment

31. A different philip on July 28, 2008 1:41 PM writes...

T:

In Process chemistry, everyone knows (or should know) exactly how long a reaction they are developing will take at a given temperature regardless of their nationality. If you're talking about graduate students, I don't know enough Japanese students to make that assertion, but I will say that coming back to grad school from an associate position in Process, people in grad school are generally less meticulous than I'm used to, but I don't blame them. You're busy enough as it is, if you have time to watch every single reaction every half hour all day then you're not setting up enough experiments. The focus in graduate school is totally different.

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
Gitcher SF5 Groups Right Here
Changing A Broken Science System
One and Done
The Latest Protein-Protein Compounds
Professor Fukuyama's Solvent Peaks
Novartis Gets Out of RNAi
Total Synthesis in Flow
Sweet Reason Lands On Its Face