Field of Science

The unstoppable Moore hits the immovable Eroom

Thanks to Derek I became familiar with an article in the recent issue of Nature Reviews Drug Discovery which addresses that existential question that has been asked by so many plaintive members of the scientific community; why has pharmaceutical productivity declined over the last two decades with no end to this attrition in sight?

The literature has been awash in articles discussing this topic but this piece presents one of the most perceptive and well-thought out analyses that I have recently read. The paper posits a law called "Eroom's Law", the opposite of Moore's Law, which seems to chart the regress in drug approvals and novel medicines contrary to Moore's bucolic vision of technological progress. The authors wisely avoid recommending solutions, but do cite four principal causes for the decline. Derek is planning to write a series of undoubtedly insightful posts on these causes. But here I want to list them especially for those who may not have access to the journal and discuss one of them in some detail.

The first cause is named the 'Better than The Beatles' effect. The title is self-explanatory; if every new drug that has to be developed is required to be better than its predecessor which has achieved Beatle-like medical status, then the bar for acceptance of this drug is going to be very high leading to an expensive and resource-intensive discovery process. An example would be a new drug for stomach ulcers which will have to top the outstanding success of ranitidine and omeprazole; unlikely to happen. Naturally the bar is very high for certain areas like heart disease with its statins and hypertension with its phalanx of therapies, but the downside of this fact is that it stops novel medication from ever seeing the light of day. The Better-than-The-Beatles bar is understandably lower for a disease like Alzheimer's where there are no effective existing therapies, so perhaps drug developers should focus on these areas. The same goes for orphan drugs which target rare diseases.

The second reason is the 'cautious regulator' with the title again being self-expalantory. The thalidomide disaster in the 1960s led to a body of regulatory schedules and frameworks that today severely constrain the standards for efficacy and toxicity that drugs have to meet. This is not bad in itself, except that it often leads to the elimination of potential candidates (whose efficacy and toxicity can be modulated later) very early on in the process. The stupendously crushing magnitude of the regulatory schedule is illustrated by the example of a new oncology drug, whose documentation if piled in a single stack would top the height of the Empire State Building. With this kind of regulation, scientists never tire of pointing out that many of the path breaking drugs approved in the 50s and 60s would never survive the FDA's gauntlet today. There's a lesson in there somewhere; it does not mean that every new compound should be directly tested on humans, but it does seem to suggest that maybe compounds which initially appear problematic should be allowed to compete in the race a little longer without having to pass litmus tests. It's also clear that, as with the Beatles problem, the regulatory bar is going to be lower for unmet needs and rare but important diseases. An interesting fact cited by the article is the relatively low standards for HIV drugs in the 90s which were partly a result of the intense lobbying in Washington.

The third reason cited in the article concerns the 'throw money at it' tendency. The authors don't really delve into this, partly because the problem is rather obvious; you cannot solve a complex, multifaceted puzzle like the discovery of a new drug simple by pumping in financial and human resources.

It's the fourth problem that I want to talk about. The authors call it the 'basic science-brute force' problem and it seems to point to a rather paradoxical contradiction; that the increasingly basic-science and data-driven approaches in the pharmaceutical industry over the last twenty years might have actually hampered progress.

The paradox is perhaps not as hard to understand as it looks if we realize just how complex the human body and its interactions with small molecules are. This was well-understood in the fifties and sixties and it led to the evaluation of small molecules largely through their effect on actual living systems (which these days is called phenotypic screening) instead of by validating their action at the molecular level. A promising new therapeutic would often be directly tested on a mouse; at a time when very little was known about protein structure and enzyme mechanisms, this seemed to be the reasonable thing to do. Surprisingly it was also perhaps the smart thing to do. As molecular biology, organic chemistry and crystallography provided us with new, high-throughput techniques to study the molecular mechanism of drugs, focus shifted from the somewhat risky whole-animal testing methods of the 60s to target-based approaches where you tried to decipher the interaction of drugs with their target proteins.

As the article describes, this thinking led to a kind of molecular reductionism, where optimizing the affinity of a ligand for a protein appeared to be the key to the development of a successful drug. The philosophy was only buttressed by the development of rapid molecular synthesis techniques like combinatorial chemistry. With thousands of new compounds and fantastic new ways to study their interactions at a molecular level, what could go wrong?

A lot, as it turns out. The complexity of biological systems ensures that the one target-one disease correlation more often than not fails. We now appreciate more than ever that new drugs and especially ones that target complex diseases like Alzheimer's or diabetes might be required to interact with multiple proteins for being effective. As the article notes, the advent of rational approaches and cutting-edge basic science might have led companies to industrialize and unduly systematize the wrong part of the drug discovery process - the early one. The paradigm only gathered steam with the brute-force approaches enabled by combinatorial chemistry and rapid screening of millions of compounds. The whole philosophy of finding the proverbial needle in the haystack ignored the possible utility of the haystack itself.

This misplaced systematization eliminated potentially promising compounds with multiple modes of action whose interactions could not be easily studied by traditional target-based methods. Not surprisingly, this led to compounds with nanomolar affinity and apparently promising properties often failing in clinical trials. Put more simply, the whole emphasis on target-based drug discovery and its attendant technologies might have resulted in lots of high-affinity, tight binding ligands, but few drugs.

Although the authors don't discuss it, we continue to have such misplaced beliefs today by thinking that genomics and all that it entails could help us to rapidly discover new drugs. As we constrain ourselves to accurate, narrowly defined features of biological systems, it deflects our attention from the less accurate but broader and more relevant features. The lesson here is simple; we are turning into the guy who looks for his keys under the street light only because it's easier to see there.

The authors of the article don't suggest simple solutions because they aren't any. But there is a hint of a solution in their recommendation of a new post in pharmaceutical organizations colorfully titled the Chief Dead Drug Officer (CDDO) whose sole job would be to document and analyze reasons for drug failures. Refreshingly, the authors suggest that the CDDO's renumeration could come in the form of delayed gratification a few years down the line when his analysis has been validated. It is hoped that the understanding emerging from such an analysis would lead to some simple but hopefully effective guidelines. In the context of the 'basic science-brute force' problem, the guidelines may allow us to decide when to use ultra-rational target-based approaches and when to use phenotypic screening or whole animal studies.

At least in some cases the right solution seems to be clear. For instance we have known for years that neurological drugs hit multiple targets in the brain. Fifty years of psychiatrists prescribing drugs for psychosis, depression and bipolar disorder have done nothing to hide the fact that even today we treat many psychiatric drugs as black boxes. With multiple subtypes of histamine, dopamine and serotonin receptors activated through all kinds of diverse, ill-understood mechanisms, it's clear that single target-based approaches for CNS drug discovery are going to be futile, while multiple target-based approaches are simply going to be too complicated in the near future. In this situation it's clear that phenotypic screening, animal studies, and careful observations of patient populations are the way to go in prioritizing and developing new psychiatric medication.

Ultimately the article illuminates a simple fact; we just don't understand biological systems well enough to discover drugs through a few well-defined approaches. And in the face of ignorance, both rational and "irrational" approaches are going to be valuable in their own right. As usual, knowing which ones to use when is going to be the trick.

Gilbert Stork on steaks, synthesis and more

Few chemists in the twentieth century have contributed as many important ideas to the science and art of organic synthesis as Gilbert Stork. Stork has made any number of groundbreaking and elegant contributions to the discipline, from the enamine reaction to radical chemistry to pathbreaking total syntheses like his synthesis of quinine. And from his perch at Columbia University where he has been for almost fifty years, he has emerged as one of his generation's most productive trainers of leading chemists in academia and industry. Here's a nice presentation listing his achievements.

Stork is now being celebrated on occasion of his 90th birthday by chemist and historian Jeff Seeman in Angewandte Chemie. Jeff brings us a wonderful collection of anecdotes, quotes and stories, both by Stork himself and by his friends and colleagues which include some of the twentieth century's leading organic chemists. There are also dozens of memorable photos. Unlike some of his contemporaries, Story is a rather unassuming man who has shunned the limelight, so it's a treat to hear these stories. There's lots of amusing stuff in there, from Stork's literally explosive relationship with cars to his being thrust into the unenviable situation of having to give a talk right after a stellar lecture by R. B. Woodward. For me two stories stood out.

First, a tale of steak disposal that momentarily triggered a panic attack and illustrated a nice lesson about kinetics (notwithstanding the fact that aqua regia contains hydrochloric, not sulfuric acid):

“There was this one really idiotic time. I remember I was really scared that I was going to blow up the entire Chemistry Department at the University of Wisconsin. I had a steak on the window ledge of my office. It was the winter, and I used the window ledge as a refrigerator. You obviously were not supposed to be cooking steaks in the lab, but I had a small lab where I was usually alone in there, and so I had a steak. But I also was not aware that biodegradable material is biodegradable, and this steak was clearly degraded on the window ledge. And the question was, what to do with it? And I decided to toss the steak in a hot acid bath which we used to clean up glassware. So, it's fuming nitric and sulfuric acid. It's really aqua regia in that bath, in that heavy lead dish, and the steak.


“And then, as I just had thrown it in there, and it fumed furiously and red fumes of who knows what, nitrous oxide of various kinds were being produced there. I became frantically concerned because fat is glycerides. So, I am hydrolyzing the fat to glycerin. You make nitroglycerine by taking glycerin and nitric acid and sulfuric acid, and obviously, I am going to produce a pile of nitroglycerine and blow up the entire building with my steak.


“Now, what is an interesting point there, why didn't it? And of course, the reason is kinetics. That is, the kinetics of oxidation of the glycerol at that temperature is much, much, much, I mean, infinitely faster than the cold temperature nitration of glycerin. And so the place was safe.”

And second, some reflections on the real value and utility of chemical synthesis:

“The toughest question to ask in synthetic organic chemistry after the work is done is: what have you learned? And you can have extraordinarily complex things. They look complex as hell. Maybe they have 80 asymmetric centers and maybe the answer is, [you've learned] nothing. I mean, you could have learned that humans are capable of enormous focused efforts and are capable of sticking with a problem which is extraordinarily complicated.


On the other hand, if somebody makes polyethylene, as somebody obvi- ously did, then you learn a lot, even though it will not thrill most synthetic chemists because this would be comparable to building a highway for an architect. I mean, it's important, but it's fairly dull compared to [building] the Guggenheim Museum, for instance... ”


“So something could be not terribly glamorous but extremely important, or vice versa. I think that B12 was vice versa. It's enormously complicated.”

That's a really important point he makes, and one that should define the choice of a research problem especially for a young investigator. There's not much point in attempting that 80 step synthesis using "hammer-and-tong" chemistry if it's not going to teach you much; that's also one of the reasons that people like Woodward get so much credit for synthesizing something first, since they really demonstrated that such complex synthesis was possible. On the other hand, throwing in two simple chemicals and watching them form an astonishingly intricate infinite lattice can really teach you something new. So can synthesizing a boringly repetitive polymer with novel properties.


The real deal in chemistry as in any other science is understanding, and the nature of experiments that impart real understanding changes with the evolution of chemical science. Stork's message for new researchers is clear; pick a problem that may not be glamorous, but whose solution would teach you something new. Stamina is not quite as important as creativity and discovery, and although perseverance is admirable, you don't make an important contribution just by proving that you can stick with a problem for ten years. Even though it may appear that way, science is not a marathon; it's scuba diving.


Image source

Why it's hard to explain drug discovery to physicists

I minored in physics in college, and ever since then I have had a lively interest in the subject and its history. Although initially trained as an organic chemist, part of the reason I decided to study computational and theoretical chemistry is because of their connections to physics by way of quantum chemistry, electrostatics and statistical thermodynamics. No other science can boast the kind of fundamental insights into the most basic workings of the universe that physics has provided in the twentieth century. Even today when we think of the purest, most exalted science we think of physics.

Not surprisingly, I have several physicist friends with whom I often talk shop. It's enormously interesting to hear about their work ranging from cosmology to solid-state physics. Yet I find that I sometimes I have trouble explaining my own work to them. And this is certainly not because they lack the capacity to understand it. It's because the nature of drug discovery is sometimes rather alien to physicists and especially to theoretical physicists. The physicists have trouble understanding drug discovery not because it's hard but because it seems too messy, unrigorous, haphazard, subject to serendipity.

But drug discovery and design is indeed all this and more, and that's precisely why it works. Success in drug discovery demands a diverse mix of skills that range from highly rigorous analysis to statistical extrapolation, gut feeling and intuition, and of course, a healthy dose of good luck. All of these are an essential part of the cocktail (to borrow a drug metaphor). No wonder that models play an integral role in the discovery of new drugs. In this sense drug discovery is very much like chemistry which Roald Hoffmann has trouble explaining to physicists for similar reasons. For a theoretical physicist, anything that cannot be accurately expressed as a differential equation subject to numerical if not analytical solution is suspect. True success in physics is exemplified by quantum electrodynamics, the most accurate theory that we know which agrees with experiment to 12 decimal places. While not as stunningly accurate as QED, most of theoretical physics in the twentieth century consisted of rigorously solving equations and getting answers that agreed with experiment to an unprecedented degree. The goal of many physicists was, and still is, to find three laws that account for at least 99% of the universe. But the situation in drug discovery is more akin to the situation in finance described by the physicist-turned-financial modeler Emanuel Derman; we drug hunters would consider ourselves lucky to find 99 laws that describe 3% of the drug discovery universe.

Physics strives to find universal laws, drug discovery thrives on exceptions. While there certainly are general principles dictating the binding of a drug to its target protein, every protein-drug system is like a human being, presenting its own quirky personality and peculiar traits that we have to deconvolute by using every tool at our disposal, whether rigorous or not. In fact as anyone in the field would know, drug discovery scientists take great satisfaction in understanding these unique details, knowing what makes that particular molecule and that particular protein tick. Try to convince any scientist working in drug discovery that you have found an equation that would allow you to predict the potency, selectivity and side-effects of a drug starting from its chemical structure and which would be universally applicable to any drug and any protein, and you will be met with ridicule.

Physicists also have to understand that in drug discovery, understanding is much more important than accuracy. There's very little point in calculating or measuring binding affinity to four decimal places, but calculating relative trends in binding affinity could be very useful, even if there are errors in the individual numbers. Far more important than calculation however is in explaining why; why a small change in a drug causes a large change in its activity, why one enantiomer causes side-effects while another does not, why making a molecule mimicking the natural substrate of a protein failed, why adding that fluorine adversely affected solubility. "Why" in turn can lead to "what I should make next", which is really what a drug hunter wants to know. In most of these cases the number of variables is so large that calculation would be hopelessly impossible in any case, but even if it were possible, dissecting every factor quantitatively is not half as important as explanation. And here's the key point; the explanation can come from any quarter and from any method of inquiry, from calculation to intuition.

This brings us to reductionism which we have discussed on this blog before. Part of the reason drug discovery can be challenging to physicists is because they are steeped in a culture of reductionism. Reductionism is the great legacy of twentieth-century physics, but while it worked spectacularly well for particle physics it doesn't quite work for drug design. A physicist may see the human body or even a protein-drug system as a complex machine whose understandings we can completely understand once we break it down into its constituent parts. But the chemical and biological systems that drug discoverers deal with are classic examples of emergent phenomena. A network of proteins displays properties that are not obvious from the behavior of the individual proteins. An aggregate of neurons displays behavior that completely belies the apparent simplicity of neuronal structure and firing. At every level there are fundamental laws governing a particular system which we have to understand. Reductionism certainly doesn't work in drug discovery in practice since the systems are so horrendously complicated, but it may not even work in principle. Physicists need to understand that drug discovery presents reductionism in a straitjacket; it can help you a little bit at every level, but it has very little wiggle room beyond that level.

Physicists may also sometimes find themselves bewildered by the inherently multidisciplinary nature of pharmaceutical research. It is impossible to discover a new drug without the contribution of people from a variety of different fields, and no one scientist can usually claim the credit for a novel therapeutic. This concept is somewhat alien especially to theoretical physicists who are used to sitting in a room with pencil and paper and uncovering the great mysteries of the universe. To be sure, there are areas of physics like experimental particle physics which now require enormous team effort (with the LHC being the ultimate incarnation of such teamwork), but even in those cases the scientists involved have been mostly physicists.

So are physicists doomed to look at drug discoverers with a jaundiced eye? I don't think so. The nature of physics itself has significantly changed in the last thirty years or so. New fields of inquiry have presented physicists with the kind of complex systems opaque to first-principles approaches that chemists and biologists are familiar with. This is apparent in disciplines like biophysics, nonlinear dynamics, atmospheric physics, and the physics of large disordered systems. Many phenomena that physicists study today, from clouds to strange new materials are complex phenomena that don't succumb to reductionist approaches. In fact, as the physicist Philip Anderson reminds us, reductionism does not even help us fully understand well-known properties like superconductivity.

The new fields demand new approaches and their complexity means that physicists have to abandon strictly first-principles approaches and indulge in the kind of modeling familiar to chemists and biologists. Even cosmology is now steeped in model-building due to the sheer complexity of the events it studies. In addition, physicists are now often required to build bridges with other disciplines. Fields like biophysics are often as interdisciplinary as anything found in drug discovery. And just like in drug discovery, physicists now have to accept the fact that a novel solution to their problem may come from a non-physicist.

All this can only be a good augury if it means that more physicists are going to join the working ranks of drug discoverers. And it will all work out splendidly as long as they are willing to occasionally hang their reductionist hats at the door, supply pragmatic solutions and not insist on getting answers right to twelve decimal places.


Image source

A new way to look for life on other planets

One of the fundamental properties of light is its polarization which refers to the spatial orientation of the electric and magnetic fields constituting a light wave. There are many fascinating facts about polarized light which are of both academic and applied interest, but perhaps the most intriguing one is the ability of chiral or handed organic compounds to change the plane of polarization of circularly polarized light. This has proven to be an invaluable tool for chemists in detecting and assigning the structures of molecules, especially biological molecules like sugars and amino acids which tend to exist in only form (left or right handed) in nature.

This fact has now been put to good use by a team of Chilean, Spanish and British astronomers who, in a paper in this week's Nature, have come up with a novel way to detect biosignatures of life on planets. They demonstrate their method by detecting the polarization signatures of water, clouds and vegetation in Earthshine. Earthshine refers to the sunlight reflected by the earth which has been reflected back by the moon towards the earth. It turns out that earthshine contains polarized light whose polarization has been shaped by the earth's atmosphere and vegetation and their constituent molecules. Specific molecules polarize specific wavelengths of light so scanning the whole range of wavelengths is essential. Crucially, the presence of vegetation is manifested in the polarization of the light by chlorophyll. Chlorophyll is special because it absorbs light up to about 700 nm. Beyond 700 nm (in the infrared region) it sharply reflects it, leading to a spike in the spectrum known as the "red edge". That is why plants glow strongly in infrared light. The red edge is a major part of the earth's reflected light as detected in outer space. It's a remarkable phenomenon which could be put to good use to detect similar life-enabling pigments on other planets.

The team used the Very Large Telescope in Chile to analyze reflected earthshine during two months, April and June. The two different times were necessary to make observations of two different viewing faces which the earth presented to the moon; one face was predominantly covered by land and vegetation and the other mainly by water. The earthshine arising from the two faces would be characterized by different spectroscopic signatures, one belonging mainly to vegetation and the other to water. At each wavelength of light they observed peaks and discontinuities corresponding to oxygen, water vapor and chlorophyll. They then compared these observations to calculations from a model that contains as parameters varying proportions of vegetation and ocean. There is some uncertainty because of assumptions about cloud structure but overall there is good agreement. Remarkably, the vegetation is sensitive to even a 10% difference in vegetation.

The technique is fascinating and promises to be useful in being able to make gross detections of water, oxygen and vegetation on other earth-like planets, all of which are strong indicators of life. Yet it is clear that earthshine presents a relatively simple test case, mainly because of the proximity of the moon which is the source of the polarized light. By astronomical distances the moon is right next to the earth and there's very little in the intervening medium by way of dust, ice and other celestial bodies. The situation is going to be quite different for detecting polarized reflections from planets that are millions of light years away. A few thoughts and questions:

1. The authors note that the lunar surface partially depolarizes the light. Wouldn't this happen much more with light coming from very far that has hit multiple potentially depolarizing surfaces? Light could also be depolarized by dense atmospheres or by interstellar media like dust grains and ice grains. More interestingly, the polarization could also be reversed or affected by chiral compounds in outer space.


2. A related question: how intense does the light have to be when it reaches the detectors? Presumably light from worlds that are billions of light years away is going to strongly interact with surfaces and interstellar media and lose most of its intensity.


3. It's clear that chlorophyll is responsible for the signature of vegetation. Alien plants may not necessarily utilize chlorophyll as the light harvesting pigment, in fact they may well be equipped to use alternative wavelengths. There could also be life not dependent on sunlight. How we will be able to interpret signatures arising from other unknown pigments and constituents of life is an open question.


4. It is likely that advanced civilizations have discovered this method of detecting life. Could they be deliberately broadcasting polarized light to signal their presence? In the spirit of a past post, could they do this with specific molecules like amino acids, isotopically labeled molecules or stereoisomers? How sensitive is the polarization to molecular concentration? Any of these compounds would strongly suggest the presence of intelligent life which has developed the technology for the synthesis and purification of organic molecules.


Image source


Sterzik, M., Bagnulo, S., & Palle, E. (2012). Biosignatures as revealed by spectropolarimetry of Earthshine Nature, 483 (7387), 64-66 DOI: 10.1038/nature10778

The right questions about modeling in drug discovery

There's been a fair bit of discussion recently about the role of modeling in drug design, and as a modeler I have found it encouraging to see people recognize the roles that modeling to varying extents plays in the whole "rational" drug design paradigm. But sometimes I also see a question being asked that's not quite how it should be pitched.

This question is usually along the lines of "Would you (as a medicinal chemist) or would you not make a compound that a modeler recommends?". Firstly, the way the question is set up ignores the statistical process evident in all kinds of drug design and not just modeling. You never recommend a single compound or even five of them but usually a series, often built around a common scaffold. Now suppose the chemist makes fifty compounds and finds only one of them to be active. Does this mean the modeling was useful? It depends. If the one hit that you get is a novel scaffold, or a compound with better properties, or one without that pesky unstable ester, or even a compound similar to ones that you have made that nonetheless picks up some hitherto unexplored interactions with the protein, then modeling would not have been in vain in spite of what is technically a low hit rate. The purpose of modeling is to provide interesting ideas, not rock-solid predictions. Plus you always have to ask the other question: "How likely is it that you would have come up with this structure without the aid of modeling or structural information"? As with almost everything else in drug discovery, the value of a particular approach is not absolute and you have to judge it in the context of alternatives.

The other problem I have is that it's frequently just not possible to objectively quantify the value of modeling or any specific approach for that matter in drug design. That's because anyone who has contributed to the design of a successful drug or even a potent ligand knows how haphazard and multifaceted the process is, with contributions flying in from every direction in many guises. Every project that I have contributed to which has resulted in more potent ligands has involved a patchwork process with an essential back-and-forth exchange with the chemists. Typically I will recommend a particular compound based on some modeling metric and then the synthetic chemist will modify parts of it based on synthetic accessibility, cost of building blocks, ease of purification etc. I might then remodel it and preserve or extend the synthetic modifications.

But it can also be the other way round. The synthetic chemist will recommend a scaffold and I will suggest tweaks based on what I see on the computer, which will lead to another round of suggestions and counter-suggestions. The creative process goes both ways and the point is that it's hard to put a number on the individual contributions that go into the making of the final product. Sometimes there may be a case where one individual recommends that magic fluorine which makes all the difference, but this is usually the exception rather than the rule and more importantly, it's often not possible to trace that suggestion to a strictly rational thought process. Finally, it's key to realize that my suggestions themselves will very often be based as much on my knowledge of physical and organic chemistry as on any specific modeling technique. Thus, even internally it's not always possible for me to separate modeling-based suggestions from other kinds. On a related note, the frequent necessity of this interactive process is why I tend to find the "modeling vs medicinal chemistry" turf wars a little irksome (although they sometimes do help in fleshing out the details); the fact of the matter is that both kinds of scientists are ultimately an important part of successful drug discovery projects involving any kind of model-building.

Now this rather unmethodical approach makes it difficult to pin down the role of any specific technology in the drug discovery process, and this usually makes me skeptical every time someone writes a paper trying to assess "The contribution of HTS/virtual screening/fragment-based drug design in the discovery of major drugs during the last decade". But that's how drug discovery is, a hodgepodge of recommendations and tweaks made by a variety of rational and irrational human beings in an environment where any kind of idea from any corner is welcome.

The man who made it possible

There's a nice set of articles in this week's Nature celebrating the birth centenary and work of a man whose work underlies almost all of modern life - Alan Turing. Considering the complete transformation of human life that computing has enabled, Turing's work along with that of his fellow computer pioneer John von Neumann will likely be recognized as one of those watersheds in human history, comparable to the invention of the calculus and the discovery of electricity. It is remarkable to consider how many millions of software engineers and billions of dollars in revenue are being generated every day off these scientists' ideas.

The historian and writer George Dyson starts off by documenting Turing and von Neumann's basic achievements. Dyson has a book out next week on this very topic, and given his past writings on technology and computing I am looking forward to it. As Dyson tells us, the basic groundbreaking idea of Turing and von Neumann was not just a machine which performs calculations at superhuman speed. It was the concept of a shared program and the then startling idea that you could code both the data and the instructions for it in the same language (binary) in the same machine. This idea really is one of those "big ideas", simple to state but absolutely revolutionary in its impact. Turing and von Neumann's greatness thus lies not in conceiving the physical manifestation (or 'instantiation' as programmers would say) of a computing recipe but in their abstract generalization of the very idea of a programmed computer.

What is not always recognized is that Von Neumann went a step ahead and floated an even more remarkable notion, that of a machine which contains instructions for assembling copies of itself. Von Neumann immediately tied this to biology; but this was a few years before Watson and Crick discovered the specific mechanism in the form of DNA base pairing. Unlike the "dynamic duo", nobody remembers von Neumann as making a signal conceptual contribution to biology. I remember the writer John Casti lamenting the general public's lack of recognition of von Neumann as the man who first really thought of the mechanism of heredity on the general basis that mathematicians are used to. As Casti pithily put it in his wonderful book 'Paradigms Lost': "Such are the fruits of the theoretician, especially one who solves 'only' the general case". To be fair, biology is an experimental science and no amount of theorizing can nail an experimental fact, but I suspect mathematicians would widely commiserate with Casti's lament.

The biologist Sydney Brenner then follows up by recognizing Turing's contributions to biology. In 1952 he wrote what is considered the first paper on nonlinear dynamics in which he described pattern formation in chemical reactions and possibly in developmental biology. We are still trying to understand the implications of that discovery even as nonlinear dynamics itself has become an enormously fruitful field with deep applications in modeling almost any complex natural phenomenon.

Finally, mathematician Barry Cooper from the University of Leeds points out a tantalizing question stemming from Turing's work; are complex, emergent phenomena strictly computable? This is partly a question about the limits of 'strong' reductionism and one that I have explored on this blog often. We don't yet know the answer to this question, to whether we can compute higher-order phenomena starting from a few simple particles and fields bequeathed to us by the particle physicists. As we tackle problems as complex as the future of the cosmos and the source of consciousness in the twenty-first century, this question will continue to hound scientists and philosophers. It waits for an answer as profound as Turing's answer to Hilbert's famous Entscheidungsproblem.

The end of biofuels?

Hartmut Michel from the Max Planck Institute of Biophysics has an editorial (open access!) in Angewandte Chemie with a title that makes his views clear: "The Nonsense of Biofuels". He essentially comes down hard on biofuel proponents of all stripes and not just the much hyped ethanol-from-corn lobby. Michel won the Nobel Prize for cracking open the structure of one of earth's most important proteins - the photosynthetic reaction center - so he certainly knows his photosynthesis.

He starts by looking at the energy efficiency of the process. It's not always appreciated that for all its rightly deserved glory, photosynthesis is not as efficient as we think, which would indeed be the case for something that's been tuned by evolutionary fits and starts and historical contingency. For one thing, UV, IR and green light cannot be utilized by plants so that leaves out a pretty high-energy part of the spectrum. Then there's all the wonderful machinery of electron transfer and light-harvesting proteins involved in the dark and light processes. The light step essentially captures photon energy and generates NADPH and ATP, and the dark step uses this energy source and reducing potential to synthesize carbohydrates from CO2. Considering the inefficiencies inherent in using the energy of massless, transient photons, only about 12% of energy from sunlight is stored as NADPH.

Then there's the question of light intensity which seems to invoke a classic catch-22 situation. At low intensities where the process is most efficient you don't have a lot of photons by definition. But try to improve the efficiency by bumping up the intensity and you get photodamage which, in Michel's words, 3.5 billion years of evolution hasn't been able to circumvent. To avoid this photodamage, plants have to recycle one of the key proteins in photosystem II about thrice every hour which inherently limits the efficiency. Finally, the key protein in the second step, RuBisCO, has a hard time distinguishing between CO2 and O2. A significant amount of energy has to be spent in getting rid of the product formed from O2 insertion.

All these hurdles lead to a rather drastic lowering of photosynthetic efficiency which gets watered down to a rather measly (but still staggeringly efficient by human standards) 4% or so.

It's pretty clear from this description that any kind of efforts to get better efficiency from biofuels will have to overcome enormous protein engineering hurdles. This does not bode well for current studies aimed at such goals. While these are very exciting from an academic standpoint, they will have to lead to a very drastic retooling of the basic photosynthetic apparatus, involving re-engineering numerous genetic pathways and their products, to be of large-scale commercial value. It's all too easy to underestimate the sheer amount of energy that we want to generate from these technologies. I feel the same about the synthetic biology efforts that seek to produce all kinds of valuable industrial chemicals and drugs from engineering bacteria. These efforts are undoubtedly promising, but getting bacteria to do something which they have not evolved to and that too on a scale rivaling the fossil fuel industry is a very long shot indeed. Michel doesn't even seem optimistic about the recent excitement regarding biofuel production from red algae, and reading his prognosis one wonders how much collaborations such as the one between Exxon and Craig Venter are actually going to yield. And finally, there's all the alternatives which that land for biofuels and the biofuel feedstock itself can be put to, which is a question still being pondered.

Michel is much more optimistic about photovoltaics which already promise energy conversion efficiencies of more than 15%. When the end product is used in a car, photovoltaic batteries also funnel about 80% of their energy into propelling the vehicle. In addition, Michel sees promise in recent advances in battery technology leading to much higher energy density.

Personally I am a proponent of context-specific energy use. I think that considering the vast variation of resource distribution, geography, energy requirements, paying capacity and economics, it doesn't make much sense to search for any one-size-fits-all solution. But that seems to usually be the case every time someone touts a single, seemingly miraculous solution as being universally applicable around the world. I have similar thoughts about solar energy. The solutions currently available don't seem to solve the problem of transmission and availability in regions where the sun doesn't shine that much. Part of Michel's "vision" is the widespread deployment of superconducting cables which even now (more than 25 years after the discovery of "high-temperature" superconductors) seems like a minor fantasy. Notwithstanding these issues though, solar power certainly seems to have a much larger role to play in our economy than it currently does, especially in regions which get plenty of sunlight.

But biofuels? The problems there seem much more grounded in fundamental biological constraints rather than technological ones. And it's hard to overturn 3.5 billion years of evolution so I am not sure I should hold my breath. Time will tell.

Identifying new drug targets in zebrafish

Classically, when very little about molecular biology and protein structure was known, one of the best methods to discover drugs was to try to guess physiological effects by looking at drug similarity. The rationale was simple; drug A has this chemical structure and lowers blood pressure. Drug B which is being used for a completely different indication also has a similar chemical structure. Maybe we can use drug B for lowering blood pressure too?

In the absence of rational approaches this technique (which these days is called "repurposing") could be surprisingly fruitful. However, as molecular biology, crystallography and structure-based drug design took off in the 80s, rational drug discovery became much more focused on protein structure and drug developers began trying to guess drug function by looking at target similarity in terms of sequences and binding pockets.

But the relative lack of dividends from structure-based drug design (which nonetheless continues to be important) has led to a re-examination of the old paradigm. One of the most interesting advances to come out of this thinking was pioneered by Brian Shoichet and Bryan Roth's groups (at UCSF and UNC-Chapel Hill) a few years back. Their rationale was simple too; look at drug similarity, albeit using modern computational methods and metrics, and try to cross-correlate drug targets using these similarity metrics. Similar drugs should hit each other's targets. The method seems somewhat crude but has worked surprisingly well. It was even listed as one of Wired magazine's top ten scientific breakthroughs of 2009.

In a recent publication the authors take the method a step further and apply it to phenotypic screening which has emerged as an increasingly popular method in the pharmaceutical industry. Phenotypic screening is attractive because it bypasses having knowledge of the exact molecular target; basically you just inject different potential drugs in a test system and look at the results which are usually quantified by a phenotypic response such as a change in membrane potential, cell differentiation or even something as general as locomotion. Once a particular compound has elicited a particular response, we can start the arduous task of finding out what it's doing at a molecular level. Not surprisingly, several proteins can be responsible for a general response like locomotion and detecting all of them is non-trivial to say the least. It is for this rather involved exercise that the authors provide a remarkably simple potential (partial) solution.

The current study looks at phenotypic screening on the zebrafish, a favorite model organism of biologists. 14,000 molecule were screened to elicit a photomotor response in the zebrafish embryos. Out of these, about 700 were deemed active. To find out the targets for these molecules, the authors interrogated their chemical "similarity" against a large group of compounds for which targets are known. Importantly, the authors use a statistical technique to calculate an expectation value (E value) which indicates whether the similarity arises by chance alone. A lower E value means a higher likelihood of statistical significance. One of the most remarkable things in these studies is that the metric for similarity is a simple, computationally cheap 2D metric called a fingerprint which looks at the presence or absence of certain functional groups in the molecules. That such a metric can work at all is remarkable because we know that an accurate estimation of similarity should ideally include the 3D conformation that the drug presents to the protein target.

Nonetheless, 284 molecules predicted to be active on 404 targets were picked based on their low E values. Out of these, 20 molecules were especially interesting because they were seen to have novel drug-target associations not seen before. When the authors tested these molecules against the relevant targets, 11 among them had activities ranging from low nanomolar to 10 µM. For a computational method this hit rate is quite impressive, although a more realistic measure of the hit rate would have come from testing all 284 compounds. The activity of the molecules was validated by running comparisons with molecules that were known to elicit the same response from the predicted targets. Confirmation also came from competition experiments with antagonists. Some of the unexpected targets predicted include the beta-adrenergic receptor, dopamine receptors and potassium ion channels.

I find these studies very encouraging. For one thing, the computational method can potentially save a huge amount of time needed to experimentally uncover the target for every active compound. As mentioned above, it's also remarkable that a very simple metric of 2D similarity yields useful results. The success rate is impressive; however, an even lower rate would still have been worth the modest effort if it resulted in new drug-target relationships (which are far more useful than new chemically similar ligands for the same target). However I do think it would have been very interesting to look at the failures. An obvious source of failure comes from using the wrong measure of similarity; at least some compounds are failing presumably because their 2D similarity does not translate to the 3D similarity in conformation required for binding to the protein target. In addition there could be protein flexibility which would result in very different binding for supposedly similar compounds. Medicinal chemists are well aware of "activity cliffs" where small differences in chemical structure lead to great differences in activity. These cliffs could also lead to lack of binding to a predicted target.

Nevertheless, in an age when drug discovery is only getting harder and phenotypic screening seems to be an increasingly important technique, these computational approaches promise to be a useful complement. Gratifyingly, the authors have developed a website where the algorithm is available for free. The technique has also spawned a startup.

Image source

Roald Hoffmann on the futility of classifying chemists

Roald Hoffmann has an editorial (open access!) in Angewandte Chemie in which he (mostly) gently scolds those who have criticized many of the last decade's Nobel Prizes as being "insufficiently chemical". I agree with him that any kind of preconceived expectations about who should get the Nobel Prize tries to fit chemistry into a straitjacket and denies scientists who may not have been trained in traditional chemistry departments the right to call themselves chemists.

As I have written elsewhere, it's partly the changing nature of what's considered important in chemical research that has shaped the face of the chemistry Nobel Prize since it was first awarded. With biology being the most exciting science of the twenty-first century and chemistry playing a foundational role in its progress, it is inevitable that more biologists are going to get chemistry prizes. And for those who may be uncomfortable with the prize awarded to biology-oriented research in the last decade, Hoffmann's observation that biology has been recognized much less over the last thirty years may provide some solace.

But any such qualms are beside the point. As Hoffmann says, the variety of chemistry Nobels given out over the years simply demonstrates the sheer reach of chemistry into multiple fields of biology, physics and even engineering. As we enter the second decade of the new century there's little doubt that fields traditionally associated with physics or engineering may increasingly be recognized by all kinds of chemistry prizes.

"Ubiquitin and the ribosome, fluorescent proteins and ion channels are as fundamentally chemical as metal surfaces, enantioselective catalysts, olefin metathesis, or, just to name some fields squarely in our profession that should be (or should have been) recognized, laser chemistry, metal–metal multiple bonding, bioinorganic chemistry, oral contraception, and green or sustainable chemistry."


And ultimately he emphasizes something that we should all constantly remind each other. It's a prize, awarded by human beings. It's an honor all right, but it does very little to highlight the objective value of the research which is usually evident far before the actual recognition. The fact that we were informally nominating Robert Grubbs or Roger Tsien years before they received the prizes makes it clear that no prize was really going to change our perception of how important their work was. Today we look at Tsien's research on green fluorescent protein with the same joyful interest that we did ten years ago.


Hoffmann sees the principal function of the Nobel Prize as providing an incentive for young students and researchers from scientifically underprivileged countries, and he cites the examples of Kenichi Fukui and Ahmed Zewail inspiring their fellow countrymen. The Nobel Prize certainly serves this function, but I have always been a little wary of pitching the benefits of scientific research by citing any kind of prize. The fact is that most people who do interesting research will never win the Nobel Prize and this does nothing to undervalue the importance of their work. So even from a strictly statistical standpoint, it would continue to be much more fruitful to point out the real benefits of science to young people- as a means of understanding the world and having fun while you are at it. Prizes may or may not follow.


Hat tip: Excimer


Image source

Book review: Philip Anderson's "More and Different"

The Nobel Prize-winning physicist Philip Warren Anderson is one of those rare species - a scientist who is not only world-class in his own field but who seems capable of saying something interesting about virtually every topic under the sun. His career at Bell Labs overlapped with the lab's most illustrious period and apart from his prizewinning work in solid-state physics, Anderson has made groundbreaking contributions to at least two other diverse fields - particle physics (he was actually the first one to suggest the existence of the Higgs boson) and the epistemology of science. In this book he holds forth on a wide variety of subjects ranging from postmodernism to superconductivity. The chapters consist of book reviews, commemorative essays, transcripts of talks, opinion pieces and a variety of other writings over the past five decades. In every chapter there are at least a few rather deep statements which deserve close scrutiny.

The book is roughly divided into three parts. The first part details Anderson's views on the history and philosophy of science including his own field - solid-state physics. The second part talks about Anderson's reminiscences and thoughts on his scientific peers, mostly in the form of book reviews that he has written for various magazines and newspapers. The third part deals with science policy and politics and the fourth is dedicated to "attempts" at popularizing science.

Some of the chapters are full of scientific details and can be best appreciated by physicists but there's also a lot of fodder for the layman in here. A running thread through several essays is Anderson's criticism of ultra-reductionism in science which is reflected in the title of the book, "More and Different". Anderson's basic viewpoint is that more is not just quantitatively but qualitatively different from less. In 1972 he made a splash by discussing in an article in Science magazine how "higher-level" sciences are based on their own fundamental laws which cannot be reduced to physics. In the book he details this philosophy through several examples from physics, chemistry, biology and psychology. He does not deny the great value of reductionism in the development of modern science but he incisively explores its limits.

Other chapters contain critiques of the latest fads in physics including string theory. Anderson bemoans string theory's lack of connection to concrete experiment and its failure to predict unique, robust solutions. He makes it clear that string theory is really mathematics and that it fails to adhere to the tried and tested philosophy of science which has been successful for almost five hundred years. Other chapters have insightful commentary on the role of mathematics in physics, Bayesian probability and physics at Bell Labs. A particularly amusing essay critiquing the current funding situation in the United States proposes a hypothetical alternative history of quantum mechanics in the US, where scientific pioneers like Dirac and Heisenberg may not have been able to do groundbreaking research because of the publish-or-perish environment and the dominance of the old guard.

There's also some valuable material in here about the sociology of science. This is exemplified by an especially insightful and detailed chapter on scientific fraud where Anderson explores the reasons why some scientists commit fraud and others don't expose it as widely as they should. In Anderson's opinion the most valuable method to expose fraud is to ask whether it destroys what he calls the "seamless web of science" - the existing framework of fundamental laws and principles that allow relatively little room for revolutionary breakthroughs on a regular basis. In many cases the web's integrity is clearly not consistent with the new finding, and the rare case where the web can subsume the new discovery and still stay intact leads us into genuinely new scientific territory. He also takes scientists to task for failing to point out the destruction of this seamless web by apparently far-reaching but fundamentally flawed new discoveries. In other chapters Anderson also comes down hard on the postmodernism distortion of science, critiquing such philosophers as Nancy Cartwright and upholding the views of debunkers like Alan Sokal. He also has some valuable commentary on science policy, especially on Star Wars and missile defense. Other writers have written much more detailed critiques of such programs, but Anderson succinctly demonstrates the limitations of the concept using commonsense thinking (The bottom line: Decoys can easily foil the system and a marginal improvement by the offense will result in a vastly increased cost for the defense).

Finally, the book contains mini sketches of some of Anderson's peers who happened to be some of the great scientific minds of the twentieth century. Anderson reviews books by and about Richard Feynman, Murray Gell-Mann, Stuart Kauffman, Stephen Hawking, Roger Penrose, John Bardeen and William Shockley among others. I happen to agree with him that books by scientists like Hawking, Penrose and Greene, while fascinating to read, paint a rather biased picture of physics and science. For one thing, they usually oversell the whole reductionist methodology in their constant drive to advertise the "Theory of Everything". But more importantly, they make it sound like particle physics and cosmology are the only games in town worth thinking about and that everything else in physics is done on the periphery. This is just not true. As Anderson makes it clear, there are lots of fields of physics including condensed matter physics, biophysics and non-linear dynamics which contain questions as exciting, fundamental and research-worthy as anything else in science. As just one example, classical physics was considered a staid old backwater of the physics world until chaos burst upon the scene. It's also clear, as was the case with chaos, that some of the most exciting advances will come from non-physicists. There are foundational phenomena and rich dividends to be mined from the intersection of physics with other fields in the twenty-first century.

Anderson's book might precisely be the kind of writing ignored by the public because they are too taken with the Hawkings, Greenes and Randalls. To those folks this volume would be an essential and healthy antidote. There's something in there for everyone, and it makes it clear that science still presents infinite horizons on every level. After all, more is different.