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

« The NIH Goes For the Gusto | Main | Insider Trading at the FDA »

March 29, 2011

Modeling and Structure

Email This Entry

Posted by Derek

Man, am I getting all kinds of comments (here and by e-mail) about my views on modeling, QSAR, and the like. I thought it might be helpful for me to clarify my position on these things.

First off, structure. It's a valuable thing to have. My comments on the recent Nature Reviews Drug Discovery article were not meant to suggest otherwise, just to point out that the set of examples the authors picked to make this point was (in my view) flawed. It's actually surprisingly hard to come up with good comparison sets that isolate the effect of having structural information on the success of drug discovery projects. There are too many variables, and too many of them aren't independent. But just because a question (does having structural information help, overall?) is hard to answer doesn't mean that the answer is "no".

As an aside, since I've talked here about my admiration for fragment-based approaches, my own opinion should have been pretty clear already. Doing fragment-based drug discovery without good structural information looks to be very hard indeed.

Now, that said, there's structure and there's structure. Like every other tool in our kit, this one can be used well or used poorly. I think that fragment projects (to pick one example) get a lot of bang-for-the-buck out of structural data, and at the opposite end of the scale are those projects that only get good X-ray data after they've sent their compound to the clinic. No, wait, let me take that back. In those cases, the structure did no good, but it also did no harm. At the true opposite end of the scale are the projects where having structural data actually slowed things down. That's not frequent, but it does happen. Sometimes you have solid data, but for one reason or another the X-ray isn't corresponding to what's happening in real life. And sometimes this kicks in when medicinal chemists try to make too much out of less compelling structural data, just because it's all they have.

Now for in silico techniques. I have a similar attitude towards modeling of all kinds, but at one further remove than physical structure data. That is, I think it can be used well or used poorly, but I think that (for various reasons) the chances of using it poorly are somewhat increased. One reason is that modeling can be very hard to do well, naturally. And at the same time, tools with which to model conformations, docking, and so on are pretty widely available, which leads to a fair amount of work from people who really don't know what they're doing. Another reason is that the validity of any given model is of limited scope, as is the case with any mental construct that we have about what our molecules are doing, whether we used a software package or waved our hands around in the air. The software-package version of some binding model is more likely to have a wider range of usefulness than the hand-waving one, but they'll both break down at some point as you explore a range of compounds.

The key then is to figure out as quickly as possible if the project you're working on would be enhanced by modeling, or if such modeling would be merely ornamental, or even harmful. And that's not always easy to do. Any reasonable model is going to need a few iterations to get up to speed, generally requiring some specific compounds to be made by the chemists, and if you're running a project, you have to decide how much effort is worth spending to do that. You don't want to end up endlessly trying to refine the model, but at the same time, that model could turn out to be very useful after a few more turns of the crank. Which way to go? The same decisions apply, naturally, to the folks standing in front of the hoods, even without any modeling. How many more compounds are worth making in a given series? Would that effort be better used somewhere else? These calls are why we're paid the approximation of the big bucks.

So, while I don't think that modeling is an invariable boon to a project, neither do I think it's a waste of time. Sometimes it's one, and sometimes it's the other, and most of the time it's a mix of each - just like ideas at the bench. When modeling works, it can be a real help in sending the chemists down a productive path. On the other hand, you can certainly run a whole project with no modeling at all, just good old-fashioned analoging from the labs. It's the job of modelers to make the first possibility more likely and more attractive, and the job of the chemists and project managers to be open to that (and to be ready to emphasize or de-emphasize things as they develop).

This point of view seems reasonable to me (which is why I hold it!) But it also exposes me to complaints from people at both ends of the spectrum. I'm a lot more skeptical of in silico approaches than are many true believers, but I don't want to make the mistake of dismissing them outright.

Comments (15) + TrackBacks (0) | Category: In Silico


1. done-it-too on March 29, 2011 2:06 PM writes...

In general, structure & modelling are tools that can, among other applications, lead to hypotheses that must then be further evaluated and tested. In the drug discovery world, neither are truly valued end- games unto themselves. Like most other tools in our varied research collection, structure based-hyptheses can be helpful in contributing to more successful ultimate outcomes, or can be disasterous if applied with so much conviction and abandon wherein the hypothesis must be right even when confronted with real experimental data suggesting otherwise (yes, folks, this does happen).

Permalink to Comment

2. Glen on March 29, 2011 5:02 PM writes...

I agree with your take on the value of structural information, It's great to have, but it's a tool like any other -- it should not be the sole driver for drug design decisions. The best modelers I have worked with always temper their suggestions with a huge grain of salt, saying, "Don't read too much into this." It can help prioritize among multiple target compounds, though, no question.

Permalink to Comment

3. Curious Wavefunction on March 29, 2011 5:33 PM writes...

It's nice to hear a balanced perspective that correctly locates the value of modeling. Modeling, HTS, FBDD etc., all these are simply tools. Just like other tools, they burn you if abused and reward you if you know what you are actually doing with them. There's no use blaming modeling if the fault lies with the modelers (some of whom are, as you mentioned, people who use these tools as black boxes) and in fact there's no reason to either exaggerate or downplay the value of any of these technologies. I never understood why all this is just seen as turf wars between the eithers and the ors.

Permalink to Comment

4. Crossvalidated on March 29, 2011 5:47 PM writes...

Completely agree, modeling and QSAR always get promoted as a blackbox solution. I've read and reviewed so many crappy QSAR papers, where people seem to believe that simply using the right program instantly makes their result completely reflective of what's going on in vivo. You really need expertise in modelling just to know where the problems lie; the computer will always give you an answer its up to the modeller to decide whether its a worthwhile answer or a load of crap.

Permalink to Comment

5. JC on March 29, 2011 6:09 PM writes...

Time wasting nonsense.

Permalink to Comment

6. Commissar on March 29, 2011 6:55 PM writes...

Modelers, chemists,...all faceless nobodies who only exist as numbers on my computer screen. UPDATE: I've just checked my screen and it appears you do not exist any longer even as abstract numbers. It appears the glorious MBA's have decided to purchase late stage product on the market so yall have to eat dirt ya hear!

Permalink to Comment

7. John Harrold on March 29, 2011 7:50 PM writes...

First off, structure. It's a valuable thing to have. My comments on the recent Nature Reviews Drug Discovery article were not meant to suggest otherwise, just to point out that the set of examples the authors picked to make this point was (in my view) flawed. ... But just because a question (does having structural information help, overall?) is hard to answer doesn't mean that the answer is "no".

That's exactly how I took your previous posting. And I say this as a modeler (well systems pharmacology and PKPD --- different but yet the same). I think a healthy skepticism toward such things very important. It keeps us math folks on our toes, and the end result is generally a much better outcome.

Permalink to Comment

8. Anthony Bishop on March 30, 2011 2:50 AM writes...

"Now for in silico techniques." I was once pitching a drug discovery company to VCs - we had some in silico data which taken at face value was very positive. The VC said (and I quote); "Virtual results will get you virtual cash!"

Permalink to Comment

9. J on March 30, 2011 3:34 AM writes...

In an Industry that is as much about creativity as it is about productivity, surely any tool that provides ideas is a great asset. A bad workman blames his tools. After the ideas are on the table it's then only about ranking, testing and refining ideas. Call then hypothesis if you prefer.

If a molecular modeller can accurately model water in-silico, then you should believe eveything he says that's based on his models.

Permalink to Comment

10. Pete on March 30, 2011 8:56 AM writes...

In my experience computational chemistry techniques have broader applicability in lead identification than lead optimisation. For the computational chemist, lead identification tends to be about geometry and shape and I prefer the term ‘directed screening’ to ‘virtual screening’. However, it is worth remembering that successful directed screening requires a strong foundation in physical organic chemistry (e.g. for setting ionisation states) and will draw heavily on experimental data (e.g. imposing conformational preferences found in crystal structure databases on screening databases) even though this may not be obvious.

In lead optimisation I try to use the computational techniques to get a better understanding of molecular properties (conformational preferences, interaction potential, ionisation, tautomers) that the team can use in design or to address specific issues. In lead optimisation I have often functioned as a physical organic chemist and used the computational tools to build on measured data. Cheminformatic approaches such as matched molecular pair analysis (MMPA) can be useful for quantifying the effects of specific structural changes on physicochemical properties. On a project where we were addressing plasma protein binding (PPB) I was able to show that replacing the existing carboxylic acid with tetrazole was almost certainly going to increase PPB. The tetazole did nort get made.

I am not a great fan of descriptor based QSAR and don’t consider a model where I can’t see the training data to be of much use. A particular concern is that global QSAR models may simply be ensembles of local models.

Permalink to Comment

11. MoMo on March 30, 2011 9:13 AM writes...

QSAR? I see little to no QSAR coming from Pharma and all from academia, unless the Q stands for qualitative, then there are boatloads. Pharma has little to say in the literature on REAL QSAR as far as I can see.

So therefore their modellers do not either A) publish their data B} Know how to do it effectively and reliably.

Its probably the latter, we hired a "Modeller" from Big Pharma once, he sat around and drank coffee, put the compounds into bioactivity "BINS" then tried to tell us how important they were.

After spending millions and getting useless data from a variety of projects we now use the SG systems as door stops, and our CEO prematurely aged.

Now when you mention compuational chemistry his temples throb...

Permalink to Comment

12. Pete on March 30, 2011 9:24 AM writes...

Binning is a standard trick for making weak relationships look much stronger than they actually are. I'm amazed that editors and reviewers fall for it so often since the results are often not worth the calorific equivalent of the paper on which they are printed.

Permalink to Comment

13. HappyDog on March 30, 2011 9:55 AM writes...


There's a third explanation you didn't mention. Modelers in pharma (well, the good ones, that is), largely realize that QSAR is a retrospective tool with little use in prospective drug design. It's fundamentally an interpolative tool rather than an extrapolative tool. The problem is that to get a good QSAR model, you use the data the project has already produced. Once you start talking about novel compounds, the chances are very good they are either outside of the model's applicability domain (thus rendering the model useless), or so obvious that the model isn't really needed. Good maybe for publishing papers after the project is over and adding a line on the CV, but not much else. I used to be enamored with QSAR earlier in my career. Now if a chemist asks me to generate a QSAR model, I ask him what he really wants to learn.

Permalink to Comment

14. DrSnowboard on March 30, 2011 11:58 AM writes...

'Life must be lived forward, but can only be understood backwards.' Kierkegaard
Could have been describing QSAR.

Permalink to Comment

15. hypnos on March 30, 2011 12:48 PM writes...

"Any reasonable model is going to need a few iterations to get up to speed, generally requiring some specific compounds to be made by the chemists, and if you're running a project, you have to decide how much effort is worth spending to do that."

I absolutely agree with that. In my experience, it is very difficult to convince people of the need to make (some) compounds that were predicted inactive. Without such examples, it is very difficult to validate a model.

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