Nobelist John Pople on using theories and models the right way

John Pople was a towering figure in theoretical and computational chemistry. He contributed to several aspects of the field, meticulously consolidated those aspects into rigorous computer algorithms for the masses and received a Nobel for his efforts. The Gaussian set of programs that he pioneered is now a mainstay of almost every lab in the world which does any kind of molecular calculation at the quantum level.

On the website of Gaussian is a tribute to Pople from his former student (and president of Gaussian) Michael Frisch. Of particular interest are Pople’s views on the role of theories and models in chemistry which make for interesting contemplation, not just for chemists but really for anyone who uses theory and modeling.

• Theorists should compute what is measured, not just what is easy to calculate
• Theorists should study systems people care about, not just what is easy or inexpensive to study.

Both these points are well taken as long as one understands that it’s often important to perform calculations on ‘easy’ model systems to benchmark techniques and software (think of spherical cows…). However it’s also a key aspect of modeling that’s often lost on people who simulate more complex systems. For instance the thrust of a lot of protein-small molecule modeling is in determining or rationalizing the binding affinity between the protein and the small molecule. This is an important goal, but it’s often quite insufficient for understanding the ‘true’ binding affinity between the two components in the highly complex milieu of a cell, where other proteins, ions, cofactors and water jostle for attention with the small molecule. Thus, while modelers should indeed try to optimize the affinity of their small molecules for proteins, they should also try to understand how these calculations might translate to a more biological context.

• Models should be calibrated carefully and the results presented with scrupulous honesty about their weaknesses as well as their strengths.

This is another aspect of theory and modeling that’s often lost in the din of communicating results that seem to make sense. The importance of training sets in validating models on known systems is well-understood, although even in this case the right kind of statistics isn’t always applied to get a real sense of the model’s behavior. But one of the simpler problems with training sets is that they are often incomplete and miss essential features that are rampant among the real world’s test sets (more pithily, all real cows as far as we know are non-spherical). This is where Pople’s point about presenting the strengths and weaknesses of models applies: if you are unsure how similar the test case is to the training set, let the experimentalists know about this limitation. Pople’s admonition also speaks to the more general one about always communicating the degree of confidence in a model to the experimentalists. Often even a crude assessment of this degree can help prioritize which experiments should be done and which ones should be best assessed against cost and implementation.

• One should recognize the strengths as well as the weaknesses of other people's models and learn from them.

Here we are talking about plagiarism in the best sense of the tradition. It is key to be able to compare different methods and borrow from their strengths. But comparing methods is also important for another, more elemental reason: without proper comparison you might often be misled into thinking that your method actually works, and more importantly that it works because of a chain of causality embedded in the technique. But if a simpler method works as well as your technique, then perhaps your technique worked not because of but in spite of the causal chain that appears so logical to you. A case in point is the whole field of molecular dynamics: Ant Nicholls from OpenEye has made the argument that you can’t really trust MD as a robust and ‘real’ technique if simpler methods are giving you the answer (often faster).

• If a model is worth implementing in software, it should be implemented in a way which is both efficient and easy to use. There is no point in creating models which are not useful to other chemists.

This should be an obvious point but it isn’t always one. One of the resounding truths in the entire field of modeling and simulation is that the best techniques are the ones which are readily accessible to other scientists and provided to them cheaply or free of cost. Gaussian itself is a good example – even today a single user license is offered for about $1500. Providing user-friendly graphical interfaces seems like a trivial corollary of this principle, but it can make a world of difference for non-specialists. The Schrodinger suite is an especially good example of user-friendly GUIs. Conversely, software for which a premium was charged died a slow death, simply because very few people could use, validate and improve it.

There do seem to be some exceptions to this rule. For instance the protein-modeling program Rosetta is still rather expensive for general industrial use. More importantly, Rosetta seems to be particularly user-unfriendly and is of greatest use to descendants of David Baker’s laboratory at the University of Washington. However the program has still seen some very notable successes, largely because the Baker tribe counts hundreds of accomplished people who are still very actively developing and using the program.

Notwithstanding such exceptions though, it seems almost inevitable in the age of open source software that only the easiest to use and cheapest programs will be widespread and successful, with lesser competitors culled in a ruthlessly Darwinian process.

Image link

"The Thexperiment Cafe": Bridging theory and experiment?

Discodermolide and dictyostatin are complex, flexible molecules that bind to the protein tubulin and promote the assembly of microtubules during cell division. This mechanism, similar to that of the bestselling drug Taxol, derails the precise timing of cell division and kills cells by causing apoptosis or cell death. Since cancer is quintessentially a disease of aberrant cell division, both molecules have emerged as potentially promising anticancer agents. Discodermolide and dictyostatin are of special interest not only because of their extraordinary potency, but especially because they seem to retain that potency against cells which have become resistant to taxol.

A year ago I co-authored a J. Med. Chem. paper that proposed a protein-bound conformation for discodermolide using a combination of NMR data and molecular modeling techniques. We followed up with a paper published last week in JACS in which we applied similar techniques to dictyostatin. In a nutshell, the two studies revealed surprising and unexpected dissimilarity in the solution and protein-bound 3D conformations of the molecules; similarity which is belied by their superficial 2D structures. While dictyostatin presents a diverse family of conformations, discodermolide sustains a remarkably constant conformation in very diverse environments (solid-state, solution, and in the protein binding site) that is enforced primarily by steric factors.

I would like to describe the work in the latest paper separately, but for now I am intrigued by another aspect of the problem. In both cases we proposed protein-bound conformations of two medicinally relevant molecules, but in both cases our conformations were not unique. In case of dictyostatin there is at least one alternative proposed conformation while in case of discodermolide there are no less than two. Of course we think that our proposed conformation better satisfies the data (otherwise we wouldn't have published the papers!), but the fact is that we are now presented with a puzzle. Which of the proposed conformations is correct and what technique would best resolve the quandary? The answer is unambiguous: x-ray crystallography on dictyostatin and discodermolide bound to tubulin should tell us what the correct conformation is.

Max Perutz once said that one of the most attractive qualities of science is that there is usually only one right answer, unlike politics where the answer depends on the viewpoint. I think this example illustrates that quality. The question is well-defined. We now have several competing proposals for the protein-bound conformations of two important molecular targets, but we know that there must be only one bound conformation in the solid-state, one right answer. Which conformation among these is it? Or is it a totally different one which has slipped through the cracks? The question is important not only because it would reveal the mode of action of a potentially novel class of anticancer drugs, but also because it could be very useful to organic chemists who could then modify the structures of the drugs based on their bound conformations to improve their potency and other properties.

In case of discodermolide, one molecule, three proposed conformations. But only one true conformation to rule them all. Which is it? In one sense the gauntlet has been thrown in front of crystallographers and the goal should be tantalizing for them, especially because there is a single right answer. The task will undoubtedly be difficult. Until now only the tubulin-binding drug taxol has succumbed to x-ray crystallography while the drug epothilone has lent itself to electron diffraction. Both dictyostatin and discodermolide are flexible molecules that won't yield to protein co-crystallization easily. And yet the solution would almost certainly result in publication in a top journal and new directions for synthetic chemists. Most importantly, it would be the definitive validation of a scientific puzzle that is currently unresolved.

But this train of thought brings to my mind another idea. Wouldn't it be great if we could have an exclusive website where experimentalists post results that theorists have to explain and theorists post results that experimentalists have to validate? The interplay between theory and experiment has of course been the bedrock of science since antiquity. But all too often, the right kind of puzzle is not clearly communicated by one group to another. Sure, if you work in a particular field, you will probably be up to speed on the literature in your field. But the sheer deluge of information ensures occasional omission, and sometimes you may also be interested in potential challenges from other areas which cannot be easily communicated to you. For instance, the dictyostatin/discodermolide puzzle may be interesting to scientists who don't have anything to do with tubulin but who are simply eager to test a new structure determination method that can be applied to such complicated molecules. As we all know, solutions to scientific puzzles can emerge from unexpected corners, and scientists sometimes may find surprises from other fields that pique their curiosity. For example, the spectacular harnessing of physics-based methods in chemistry and biology is well-known.

Yet scientists in one field cannot possibly keep track of all other fields whose developments may be attractive to them. For instance a physicist who may be developing a promising new electron diffraction technique, potentially applicable to tubulin and discodermolide, is usually not going to be aware of literature in this area. In such cases, it would be tremendously useful to have a website whose express purpose is to serve as a bridge between theorists and experimentalists. The website would be divided into the traditional fields of science along with interdisciplinary sections. Every week, a theorist or experimentalist would pose a puzzle from his or her field whose unambiguous solution he or she believes would be amenable to experimental techniques. The puzzle would be tagged with the names of all possible fields to which it could be relevant. People could vote up or down a problem which they find particularly enticing and tractable. Experimentalists from different disciplines can then take a look at the problem. The right answer could come from left field, from quarters which were completely unexpected for the scientist who posed the question. There would still be some querying that would be necessary, but the specific nature of the website would necessitate far less wading through literature from other fields than what's usually required. Similarly, experimentalists could post curious, unexplained results that would tickle theorists' grey cells.

The website could perhaps be called "The Thexperiment Cafe" or something less obnoxious. It would be a place where theorists and experimentalists rendezvous and challenge each other with specific puzzles. It could bypass the usual exhaustive literature searching and serve as a rapid delivery vehicle for problems whose solutions are unambiguous (or even ambiguous!) and which could benefit members from each camp. Experimentalists and theorists could be one big, happy family. And science will always win.

Gernot Frenking is not happy...not at all

Stable is simply "able" with a "st"



Wow. This is a first for me. Three of the heavyweights in theoretical and computational chemistry have published a set of prescriptions in Angewandte Chemie for theoretical chemists claiming to have discovered new, "stable" molecules. In response, Gernot Frenking who is a well-known chemist himself has not just published a piercing and trenchant critique in reply to this article, but they actually seem to have reproduced the text of his referee's comments as a reply. This is a lively and extremely readable debate.

In an article asking for more "realism" from theory, the three heavyweights- Roald Hoffmann, Paul von Schleyer and Henry Shaefer III- have basically come up with a roster of suggestions in response to what they see as the rather flippant declarations by theoretical chemists of molecules as "stable". One of the annoying things about theoreticians is that they regularly analyze molecules and proclaim them as stable. Experimentalists then have to sweat it out for years to actually try to make these molecules. Frequently such molecules are stable under rather extreme conditions, for example in gas phase at 4 degrees kelvin. To address the animosity that experimentalists feel against such carefree theoretical predictions, the three chemists have come up with suggestions for publication.

They make some interesting points about criteria that should be satisfied when declaring molecules as stable. In fact they think that one must do away with the word "stable" and replace it by the words "viable" and "fleeting". For example for "viable" molecules, one has to be clear about the difference between thermodynamic and kinetic stability. Molecules described as viable by theoreticians must have half lives of about a day, must be isolable in condensed phases at room temperature and pressure, and must not react easily with oxygen, nitrogen and ozone (?). Molecules with more than +1 positive or negative charge must also be included with "realistic" counterions. Molecules must even be stable under conditions of some humidity. The authors then also make suggestions about reporting accuracy and precision, and about the well-known fact that theoretically reported precision cannot be more than experimentally measured precision.

If theoreticians think these suggestions are asking for too much, they have a friend in Gernot Frenking.

Frenking batters these suggestions down by basically launching two criticisms:
1. The suggestions are too obvious and well-known to be published in Angewandte Chemie
2. The suggestions are heavily biased towards experimentalists' preferences
As Frenking puts it, he expected to walk into a "gourmet restaurant", and was served a "thin soup" instead. Ouch.

I have to say that while the suggestions made by the three prominent scientists are quite sound, Frenking's points are also well-taken. He lambasts the suggestions that realistic counterions should be included in the calculation of a molecule with multiple charges; there are already molecules with multiple charges predicted to be theoretically stable which were then isolated by experiment. Ionic molecules with charges more than + or -1 are easily isolated in condensed phases. And one of the central questions Frenking asks is; why does a molecule need to be so experimentally stable in order to justify the publication of its theoretical existence. After all there are many molecules present in interstellar space which cannot be isolated under average Joe lab conditions. Under these circumstances, Frenking is of the opinion that the distinction between "viable" and "fleeting" is "eyewash" (it's the European way of euphemism)

I resoundingly agree especially with this contention, harsh as it sounds. Why should experimentalists get an easy pass? The whole point of theory is to push the boundaries of what's experimentally possible. To suggest that one should only publish a theoretical prediction if it can easily be verified by experiment is to do disservice to the frontiers of science. While I can understand the angst that an experimentalist may feel when he sees an unusual molecule stable only under extreme conditions declared by a theoretician as "stable", that's exactly the challenge experimentalists should be up to, to devise conditions under which they can observe these short-lived molecules. If they do this they are the ones who carry the day. Since stability as is well-known is a relative term anyway, why insist on calling something "stable" only if it satisfies the everyday lab conditions of the experimentalist. I believe that it is precisely by testing the extreme frontiers of stability that chemistry progresses. And this can be done only by making things hard for experimentalists, not easy. Theoreticians pushing experimentalists and vice versa is how science itself progresses, and there is no reason for either one of them to quit questioning the boundaries of the others' domain.

There are other points and criticisms worth reading, include other referee comments which endorse the article and are also quite interesting. In the end however, I cannot answer Frenking's central question; should this article have been published in Angewandte Chemie? We should leave it for readers to judge.

Roald Hoffmann, Paul von Ragué Schleyer, Henry F. Schaefer III (2008). Predicting Molecules - More Realism, Please! Angewandte Chemie International Edition, 47 (38), 7164-7167 DOI: 10.1002/anie.200801206

Gernot Frenking (2008). No Important Suggestions Angewandte Chemie International Edition, 47 (38), 7168-7169 DOI: 10.1002/anie.200802500