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

Monthly Archives

May 31, 2005

Modeling the Modelers

Email This Entry

Posted by Derek

Comes the question, in a comment to the last post: "How much value is added by computer simulations?" Arrr. That's not easy to answer. I'd say, in some cases quite a bit, but in the majority of cases, none at all. And whether the former make up for the latter is a mighty close call. I know that there are molecular modeling success stories out there, and I've been on a project that started off dramatically with a (subsequently validated) modeling prediction. But I've been on others where the modeling was a total waste of time and effort.

Those linger in my memory, and surely account for some of my sceptical feelings about modeling in drug discovery. More specifically, there are a some major intellectual mistakes that I've seen simulations lead people into. The first one is something I alluded to recently - the awful temptation to believe that if you've seen a model of your molecule docking into a model of your target protein, then you've seen your molecule docking into your target. You haven't, you know. You've just seen someone's best guess, and the odds are excellent that it's wrong. Make a few more analog compounds, and the wonderful model that explains it all is likely to take on an unhealthy spotted appearance.

Ah, but that refines the model further, you say. And so it does - until the next analog blows that one, too. Still more refined! We have to be getting close now! But you can go through cycle after cycle of this stuff, and that leads to another trap: running the project for the sake of refining the model. It's an easy one to fall into, and it can always be justified by imagining what you'd be able to do with a simulation that explained all your compounds simultaneously. But you're not going to get one of those. As far as I know, no one ever has. And while you're chasing it, you're likely as not wandering away from the real purpose of your project, which was to find a drug. Remember? Even a drug whose binding mode you don't really understand will do, you know.

And one more pitfall is the way that modeling can constrain your thinking. If you really believe you're seeing reality (the first trap, above), then you might start ruling out whole classes of potential compounds. After all, they don't fit the model - why make 'em? And that violates one of my laws of medicinal chemistry, which I think I need to assemble into a list of their own: never talk yourself out of making an easy compound. The number of drugs that have been found by tripping over them is much larger than the number that have been found by homing in on them mathematically. Until the modelers come up with some more convincing mojo, I'll stick with my stumbling style.

Comments (5) + TrackBacks (0) | Category: Drug Development

May 30, 2005

Tailfins and All

Email This Entry

Posted by Derek

As a drug discovery project goes along, different labs tend to claim different parts of the molecule to work on. They run all sorts of variations within their territory, usually keeping the rest of the molecule at some sort of agreed-on default setting or two. Likely as not, they'll find something along the way that makes things a lot better (more potent, longer-lasting in the blood, etc.)

The natural thing to do is to combine these things, to make what I've long called a greatest-hits molecule. "Let's put that acyl group that Jim likes on there, and put best of the N-aryls from Sue's lab, and over on the side chain we'll have that solubilizing group that works so well on Wei's molecule. . .can't miss!"

Actually, these things miss about as often as they hit. Rarely have I seen a project where you can mix-and-match with confidence. You have to try these combinations, but after you've started to fill out the matix, you find that your compounds act more like this: "Well, the acyl group is good, as long as you don't have a heteroaryl group over here, but if you do then you can get away with the chloro on this position, but not if you have the amine side chain, except when there's an alpha-methyl. . ."

What's going on is that your molecules probably don't have just one way of fitting into their binding pocket in the target protein. They might have two modes; they might have twelve. There's no way to be sure, and I say that with no intention of offending the molecular modelers and their computer simulations. (But hey, if the shoe can be docked onto your foot in a low-energy conformation, wear it.) Many of these binding modes are going to have mutually incompatible features, and you can make your head vibrate trying to reconcile them into a single coherent picture.

Comments (2) + TrackBacks (0) | Category: Drug Development

May 27, 2005

Compounds for the Sake of Compounds

Email This Entry

Posted by Derek

I've started my Memorial Day weekend early (thus the mid-morning posting time.)

One of the comments to the previous post mentions the "let's make these compounds because we can" attitude, and points out that this was the fallacy underlying the combichem boom (and bust.) True enough - I should have clarified my point by saying that the compounds I was recommending were much more targeted. They're related to a structural series that we know we're interested in, but we haven't made tested anything from this particular group yet.

And, truth be told, I don't mind the blue-sky let's-make-some-compounds approach, as long as it's done in moderation. Throwing some interesting structures into the screening files is never a waste of time, although there are often more pressing things to do.

I don't approve of sending in things that are poor candidates for starting off an optimization project, though. If something with a molecular weight of 1300 hits in your assay, there's often not much you can do about it. That's at least twice a reasonable molecular weight, and large compounds like that often can't be cut down to size. Their binding modes are complex, interesting, and almost impossible to deal with in any practical manner, unfortunately. Getting a handle on things like this is a longstanding problem in drug discovery, so unless you feel like solving it, you shouldn't add to it.

Similarly, anyone who sends in reactive compounds like acid chlorides deserves a whack over the head. Those things, assuming they don't fall apart in storage, will tear up most assays they're run in, and it's not like they're ever going to be drugs. Same goes for things like organotelluriums and other out-there elements. I have a fairly liberal attitude (silicon-carbon bonds are OK with me), but there's a limit. If you think someone's going to be happy when your nickel complex hits in their enzyme assay, you are not in touch with consensus reality.

The problem with the combichem boom wasn't always the underlying compounds, although some of them were stinkers (and most of them sure could have been cleaner.) I think the real trouble was how oversold the whole thing became. If you weren't buying or cranking out huge libraries, you were missing the gold rush. Vast untapped veins of drug leads were out there in those hills! Without the hype, things wouldn't have looked so bad. But hey, without the hype, most of those libraries wouldn't have been made. . .

Comments (6) + TrackBacks (0) | Category: Drug Assays

May 25, 2005

The Voice of Experience

Email This Entry

Posted by Derek

During a meeting today, some of us were making a decision about whether or not to take a look at a particular series of compounds in some assays. I spoke up, saying "Hey, if we've got 'em, why not? Never talk yourself out of something that's easy to test."

"The voice of pragmatism", said someone else, and I responded "Darn right!" It's a real temptation in this business to think that you know more than you do. The alternative, a realistic appraisal of just how lost you (and everyone else) is, can be a bit disconcerting, and that's why I think people overvalue their expertise. "Those compounds never work," "We did something like that before," "We already know what the answer is" - these are the sounds that people make when they're trying to sound wiser than they probably are.

I have to look out for these tendencies in my own work, too. I've seen enough different sorts of projects that I do have some valuable experience to draw on. But not all of it is valuable all the time, and it's very hard to know when you're being fooled by a false correlation with something that's happened before. I don't know when something is going to be orally active, and I don't know what its blood levels are going to be like. Much as I would like to, I don't know exactly how a given compound binds to its protein target, and neither, in almost every case, does anyone else. I sure can't predict toxicity, either, and it's not for lack of motivation. One of the most valuable things I can take away from all my experience is the willingness to step back and let the results sort themselves out.

Now, there are times when you really shouldn't try something. But those times occur much less frequently than you'd think, and usually for lack of time or resources. But if both of those are available, I'm up for taking a flyer on all kinds of odd stuff. Some of the best things I've ever done looked pretty weird while I was doing them, that's for sure. But I'm at the point in my career where I'm less concerned about looking like a fool, so perhaps my best work is still to come.

Comments (3) + TrackBacks (0) | Category: Life in the Drug Labs

May 24, 2005

You Figure It Out

Email This Entry

Posted by Derek

This is as much an economics post as a pharmaceutical one. I noticed this headline from, from a newspaper in Lusaka, Zambia: "Imported Pharmaceutical Products Hindering Growth." Curious, I read the whole piece, and I have to say that it makes my head hurt. Here, you judge: Says the head of Pharmanova Zambia, one of the country's only pharmceutical manufacturers,

"The government must reduce on imported products such as drugs. For example, if you import drugs from India, you are giving the Indian government access to growth of their economy and that does not benefit the Zambian economy at all," Dharkar said.

"What we want as manufacturers of pharmaceutical products is support from our government. We are capable of producing high quality drugs for the Zambian community."

Now, break that statement down. What Mr. Dharkar wants is "support" from his government - by which I assume he means high tariffs - so he can make all the money from the Zambian pharmaceutical market himself. This, as far as I can see, benefits his customers not one tiny bit, and it only benefits his employees until they need to buy some drugs themselves. And it's supposed to benefit the Zambian economy. . .how, exactly?

Look, this is not an intellectual property issue, and it's not an issue of recouping research costs (as it is in the US market.) We're not going to get our costs back from Zambia, a desperately poor country full of desperately poor people. As far as I can see, protecting this local factory will only cause these people to pay more for their medicines. And Mr. Dharkar agrees, but then he disagrees, right in the same paragraph:

Dharkar said most Zambian communities needed locally produced drugs, which he said would be more affordable than the imported drugs. He said his company was capable of producing antiretroviral drugs, although he observed that the high cost of producing the drugs would hinder their production in Zambia.

Now, if his drugs are more affordable, why isn't he wiping the imported stuff right out of the market by undercutting them? What's stopping him? And as for those antiretrovirals, which Zambia most definitely needs, if the cost of producing them locally is too high, then why not. . .import them from India? Aaargh.

Comments (2) + TrackBacks (0) | Category: Business and Markets

May 23, 2005

Intelligent Chemical Design

Email This Entry

Posted by Derek

I haven't commented on the controversy about including "Intelligent Design" in school curricula, but I don't want that to be interpreted as any kind of approval. On the contrary - until it offers some testable predictions, which would seem an unlikely thing to hope for, I don't see how ID even rises to the level of a preliminary theory, much less one that can compete with the level of evidence backing up evolution. Many of ID's advocates, to a greater or lesser degree, strike me as intellectually dishonest.

Intelligent Design proponents are fond of arguing about "irreducible complexity", the idea that some structures are too complicated to have been generated through stepwise evolution. They argue this on the anatomical level, which I don't buy, but I'm not going to debate that one in this forum. (Allow me to refer the curious to my fellow Corantean Carl Zimmer, who's had plenty of run-ins with these folks, and his fine introduction to evolution. Those interested in the latest news on the ID/evolution battles should check out The Panda's Thumb. For sheer mockery, often irresistible in these cases, try this.)

But when they start making arguments at the chemical level, the what-are-the-odds stuff about proteins and DNA, well, that's when I come out of my lair. A paper in the latest issue of the journal ChemBioChem got me thinking about this today. (If you have access to Wiley journals, it's here as a PDF.) It's an update of the analytical work still being done on the Murchison meteorite (a href="">PDF), which fell in Australia in 1969. The more than 100 kg of recovered Murchison material have been attacked over the years with just about every instrument of the constantly shifting state of the art in analytical chemistry.

Why all the interest? Well, a short answer is that the pieces of this meteorite reek. Even now, they smell like low-grade gasoline, and they had a powerful odor indeed when they were freshly collected. The Murchison fall is a wonderful example of a rare class of meteorites called carbonaceous chondrites. Many people don't realize how much organic gunk is floating around out in space, but there are surely millions of tons of this stuff wandering around our solar system alone.

What's in the Murchison pieces? The list continues to lengthen. We're up to at least 500 different soluble compounds, but much more of the material is dark polymeric asphalty stuff that's hard to analyze. Most famously, the meteorite contains many amino acids. Save glycine, those come in left- and right-handed isomers, and a major find is that the Murchison material is slightly biased toward the left-handed ones, which happen to be the ones that life on Earth is built around. This is an important point: the chemicals that life as we know it is composed of are not at all odd or unlikely. They're all over our solar system, they're in interstellar clouds, and there's every reason to think that they're smeared and splatted all over the universe.

And more of the stuff is being made all the time. In 2002, several research groups took icy mixtures of water, methanol, ammonia, HCN, carbon dioxide and carbon monoxide (just the sort of mixtures that you see in cometary ices and the above-mentioned interstellar clouds.) They irradiated them with ultraviolet light - as would come from the Sun or untold billions of other stars - at cold outer-space temperatures, and obtained over a dozen of the most common amino acids - here are some more details.

So, here's another key point: the really big step is between making random chemical combinations and having carbohydrates and amino acids as inevitable products. Believe me, the molecules of life are an infinitesmal sliver of all the possible backbones of up to ten or twelve carbon, nitrogen, and oxygen atoms. But organic chemistry, with no active hand on the controls, turns out uncountable heaps of them. Compared to that, the gaps that need to be filled in on the way to living systems don't seem so large.

This, to me, is one of the major stories of the last few decades. Starting hundreds of years ago, astronomy gradually moved the Earth out of its supposed spot in the center of the universe and placed it in the huge (and hugely strange) context of the universe that we now know. Now chemistry is moving us away from the view of life as a strange and precious anomaly - granted, perhaps, by a divine being? - to something that could be everywhere and may well start of its own accord. The building blocks are ubiquitous, and if you give them half a chance they start to stack themselves up.

For better or worse, the presence of an active Designer does not suggest itself. That may not seem right to some people, for many different reasons. But if there's one thing that science has been showing us, it's the the universe doesn't much care what we think about it.

Comments (20) + TrackBacks (0) | Category: Intelligent Design | Life As We (Don't) Know It

May 22, 2005

That Has to Be Good, Right?

Email This Entry

Posted by Derek

Here's a nice article over at Forbes which highlights a tough question: what's the best indicator of efficacy for a new cancer drug? One of the easiest things to measure is tumor shrinkage, and you'd think that would be a good sign. Common sense can only take you so far, though:

"But now doctors are finding that tumor shrinkage, on its own, isn't necessarily a good reason to use a drug, because it's entirely possible to shrink a tumor without helping the patient. More important measures of efficacy are how long it takes for the cancer to start worsening and how long patients live. If a drug increases survival time, its efficacy clearly outweighs any side effects."

Survival, of course, takes a lot longer to measure, and time is nowhere more equivalent to money than in a clinical trial. If we're going to start moving away from the classic response rate - number of patients with 30% or more shrinkage - then we're going to be spending a lot more to evaluate our drugs. It's true that the article linked above quotes someone who doesn't buy this logic, but he's talking about using some genetic marker as a surrogate. This is a very nice idea, but we're back to the same problem: it also takes years to prove that these things are linked to survival, and that'll always be the real standard. Who cares what your cancer is up to if it kills you on the same schedule?

No, using survival as the endpoint will almost always cost more. But I think it'll be worth it. I've never understood the benefit to desperate patients of giving them something that's just going to make them spend their last months being jerked around.

Comments (6) + TrackBacks (0) | Category: Cancer

May 20, 2005

Outside Reading

Email This Entry

Posted by Derek

Home and work have conspired to leave no time for a post this morning - man, I wish those two wouldn't team up so often; they're supposed to not even be friends or anything.

But you can head over to the science/medicine blogfest called Tangled Bank and find a lot of good stuff, written by a lot of good people that I need to blogroll. Enjoy, and I'll see everyone on Monday.

(Oh, one other thing on the outside-reading front. Glenn "Instapundit" Reynolds mentions that he has a law-school colleague who tried to build a Dune-style "stillsuit" some years back, and adduces this as proof of her geek credentials. Point taken, but let me pull my papers out for inspection. A more technically inclined geek would have realized the physics problem that Frank Herbert sort of, er, skipped over. The major point of sweating in hot weather is evaporative cooling. What happens when you trap that moisture and try to recycle it? Condensative heating, that's what, which you don't hear about as much but is as real as the laws of thermodynamics can make it. A stillsuit would cook you like a crock-pot.)

Comments (7) + TrackBacks (0) | Category: Blog Housekeeping

May 18, 2005

Vertex Turns Over a Winner

Email This Entry

Posted by Derek

Vertex announced some impressive clinical data against hepatitis C the other day, which has been a fine thing for their stock price. It looks to be a fine thing for people infected with hepatitis as well, with better results than the current interferon therapy in a much shorter time.

I'd have to guess that the side effects are lower, too, interferons being rather powerful things. Schering-Plough and Roche have been beating each other silly in that market for years now - these results point to a future where they might each end up with a much lower market share. (If Vertex's drug works even better in combination with them, though, everyone might still do just fine.)

Other things being equal, or even in the neighborhood of equal, a small organic molecule is going to beat a biotech protein every time. They're easier and cheaper to make, and easier to store and dispense. And as for dosing, well, you can get orally active small molecules (like Vertex's) - try getting an orally active protein, and get back to me when you do. Mind you, it looks like Vertex's compound really has to be taken in mighty quantities (750 mg t.i.d., that is, three times a day), but anti-infectives often have to be hammered in like this. It's still surely going to be cheaper than interferon therapy.

Vertex being Vertex, I'm sure that there's a nice presentation about all the contributions that molecular modeling made to this compound's development. But I've no doubt that their large bunch of capable medicinal chemists made their mark on it, too. Congratulations to the lot of them!

Comments (2) + TrackBacks (0) | Category: Infectious Diseases

May 17, 2005

Very Wrong, or Very Right

Email This Entry

Posted by Derek

It's not completely fair of me to make fun of the old hype about rational drug design, because every moment has its overhyped technology. (Perhaps, as we've speculated around here before, today's candidate is RNA interference. . .) All of it ends up sounding silly in the end.

And the arrogant tone that the proponents of some new systems often take sounds laughable, too, after things don't work out. But that same attitude is probably needed, up to a point. You really have to have some nerve to remake a scientific field. After all, at the very least you're saying to everyone that there's something important that they don't know about yet. And sometimes, the message is a flat "You people have this stuff completely wrong, so step back and let me show you why." It's not a job for the meek.

People with shy and fearful personalities will almost never make a great discoveries in the first place, much less publicize them effectively. That kind of thinking will cripple you with all the reasons why things won't work, why someone else (surely smarter and more competent!) would have already tried this, and so on. And even world-beating ideas tend to fail a lot before they finally get going, so the timid or easily discouraged will be convinced that they're wrong before they ever get a chance to be right.

I'm not saying that all the great discoverers are intolerable, although some of them sure are. But even if they're good to the people around them, they're mighty hard on nature and on their experiments, and harder still on the existing order.

Comments (6) + TrackBacks (0) | Category: Who Discovers and Why

May 16, 2005

Days of Silicon and Roses

Email This Entry

Posted by Derek

I thought that everyone might enjoy a quick trip back to the past with me. In my files, turning a bit yellow around the edges, is a copy of a story from Business Week's May 13, 1991 issue. It's "The Search For Superdrugs: A Union of Biotech and Chemistry May Conquer the Great Killers." Hot stuff, eh? This appears to have been an attempt to introduce BW's readership to the idea of rational drug design, which had been a hot topic for several years at the time. It looks like the people at Vertex were a good source for the reporters:

"Boger (of Vertex) is at the vanguard of a revolutionary approach to drug development. Few drugmakers admit it, but privately their executives say traditional research techniques just don't cut it anymore. The old way - screening thousands of chemical in a hit-or-miss search - is inefficient and wastes time. That's why it can now cost more than $200 million to bring a drug to market."

My first reaction to this is, "Ah, for the good old days, when that's all it cost us." I also note that the one-foot-in-the-grave method of random screening is still very much with us. Vertex's venture-capital pitch back then seems to have been very much into screen-bashing, with the inference that their hot new rational technology would (arrow into the bulls-eye sound effect here) zzzzzZZZISH-thunk right into the target instead.

I also note that the whole point of the paragraph is nonsense. Even back in 1991, the time and effort of random screening contributed at most a few percent to the eventual price tag of a new drug. The last two sentences could just as plausibly read "The old way of installing tires - with compressed air and lug nuts - is inefficient and wastes time. That's why it can now cost more than $30,000 to buy a new SUV." Well, let's press on:

"Boger and others like him are carrying the flag for a new wave of R&D often called "rational drug design." Dozens of these entrepreneurs are consummating the long-awaited combination of biotechnology and chemistry - a union that promises to streamline and enhance drug development and reshape the biotech and pharmaceutical industries. This year, the first "superdrugs" from this marriage will enter human testing. "Everything is coming together," says Brook Byers, a biotechnology venture capitalist. "Chemistry and Biology together will create the drugs of the '90s."

Well, I can't argue with that statement, except by pointing out that Chemistry and Biology together created the drugs of the '70s and '80s, too. (Byers is very much still with us, by the way.) But this naturally leads one to ask: what are these superdrugs from 1991, and where are they today?

One that they mention is an RNAse H inhibitor from Agouron, which was a very neat place to visit in the early 1990s. There's no telling how many journalists went out there to put on goggles and watch models of molecules docking into models of proteins. Then they went back to write their stories and generally made a severe category mistake, thinking that they had been watching actual drug molecules docking into real proteins. While Agouron found some success against HIV protease, the whole RNAse H inhibitor idea has had a tough time of it, although people are still plugging away.

Another drug the article brings up is an glycoprotein IIb/IIIa compound for cardiovascular disease from Genentech. This, I assume, is what eventually turned into sibrafiban (PDF file). Unfortunately, that whole class of drugs didn't work out too well when compared head-to-head against aspirin, and that was pretty much that.

We'll leave aside, for now, the fact that all these folks were trying to compute their way to fame and riches with 1991 hardware, because it's not any easier with the latest 2005 stuff out of the box. Those two examples show you exactly why we're not awash in those wonderful 90's drugs right now. The most important parts of drug development are not yet amenable to a rational approach. We simply don't know enough. If any of the companies developing oral IIb/IIIa ligands thought there was a good chance that they'd lose out to aspirin, of all the cheap competitors, they'd have run away screaming. But they didn't know, and the only way to find out was to spend the money and take the risk. No fancy graphics could have saved them.

Comments (5) + TrackBacks (0) | Category: Drug Industry History

Cuprate Voodoo

Email This Entry

Posted by Derek

I may talk a good game, but don't let me fool you: I really don't understand some of the things that happen in my lab. I was talking with some colleagues the other day, and told them about a reaction I had in my PhD work. I was using a so-called "higher-order cuprate" reagent, which you make from copper cyanide and lithium reagents, and it worked just fine for me the first time out.

And then it worked just fine again, on a larger scale, and then it worked just fine one more time. And at that point, it stopped working, forever. I never got the tiniest bit of product out of the reaction again, only starting material and varioius forms of junk. It was as if someone had turned a switch.

We're supposed to be able to work these problems out, and I tried. I checked my solvents, my starting materials, and my technique. I analyzed my lithium reagents, and I tried every bottle of copper cyanide we had. Then I made my own, the most beautiful copper cyanide you'd ever want, assuming that you'd ever want some. Result, zilch. I reversed field, thinking that I'd used an old bottle which might have some oxidation products in it that made things work. But aging the stuff by heating it in an open flask didn't help, either. Nothing did. I kept looking at my flask of product from before, a key intermediate in my synthesis, and wondering how on Earth I'd ever made it.

Well, as I finally wrote in my dissertations, "cuprate reagents of all sorts were abandoned after this experience." I never did find out what happened, and eventually worked out another way to get the product (which reminds me, if anyone out there needs to use dimethylmagnesium, talk to me - have I got a recipe for you!) Organometallic reagents are famous for being tricky, but you'd think that a science as old as organic chemistry - we are a science, right? - would be able to deal with such problems. Maybe not.

(Note: the structure and reactivity of all kinds of cuprate reagents have been the subject of food fights for many years. Here's a sample, but if you're not seriously into organometallic chemistry you're going to feel as if you'd stepped into a room full of people discussing Zoroastrian theology. A later and less contentious look at the field can be found here as a PDF.)

Comments (4) + TrackBacks (0) | Category: Life in the Drug Labs

May 13, 2005

ASCO Fever

Email This Entry

Posted by Derek

Well, today is the start of the ASCO meeting, which as I've mentioned is an interesting, important blizzard of hype and spin. A number of companies (Imclone, Merck KGaA, Bayer, Genentech, Pfizer and others) have presentations that will be watched closely. Some of these will take place over the weekend, which will at least keep a few stocks from having to halt trading (unless there are some big order imbalances come Monday morning, that is. . .)

It's hard to keep a proper perspective on this sort of presentation. One thing to remember is that everyone involved realizes the spotlight that they're under, and has planned accordingly. I've long thought that scientific meetings, as they've come to be run, are one of the worst places to discuss scientific results. I'm sure that many of the interesting and important conclusions are things that would only become clear after sitting down with the complete data sets for a few days (or weeks). Distilling all of it down to a meeting talk, even assuming (charitably) that you're not trying to sell something or divert attention, is going to degrade the information.

So, enjoy the news bulletins, and good luck with the stocks prices. But don't take ASCO more seriously than it deserves.

(By the way, another preview of the conference can be found here. Can anyone tell me what the author means in her last sentence, "The financial component is one area that is sorely lacking in research"? Here I thought that we were getting beaten up on for making the "financial component" too darn important. . .)

Comments (2) + TrackBacks (0) | Category: Cancer

May 11, 2005

Public Utility, You Say?

Email This Entry

Posted by Derek

I wanted to return to that idea of Merck (and the other drug companies) settling into a role as a "regulated public utility." You can find calls for that sort of thing (in Marcia Angell's writings, for example), and there seem to be people who find the whole idea really appealing.

Me, I find it disgusting. And I don't even know which part of the idea I find most revolting. Is it the thought of working for a pharmaceutical version of the Post Office, with what would surely be a dynamic risk-taking culture? Is it the thought of waiting. . .waiting. . .while higher and higher bureaucratic commissions and review boards grind slowly on to tell us what we should work on next?

Or is it just that the idea is so unworkable that I feel pity for a person who advocates it? Look, electricity and water are utilities. Rail service can be treated as one (and doesn't Amtrak do a fine job at it?) But these are mature industries whose task is to deliver steady amounts of known goods. The drug industry doesn't fit any of those criteria.

We argue about what diseases to treat, and how to attack them. We argue about which drugs to try, and how to use them. We disagree about the best ways to start finding the drugs in the first place. We don't even understand many of our therapeutic targets that well, and there are many others that we haven't even heard of yet. Different populations need our products in completely different ways at different times, and vary widely in their ability (not to mention their willingness) to pay for them. Does this sound like a good candidate for public utility theory?

Only if you imagine a big machine, grinding away and producing grey blocks of something called "Health Care." It all reminds me of what someone said about Woodrow Wilson, that he had the most abstract mind the speaker had ever encountered. A normal person might watch a train go by and see a train, he said. Wilson, though, saw Transportation.

Comments (4) + TrackBacks (0) | Category: Business and Markets

May 10, 2005

Getting Hired as a PhD

Email This Entry

Posted by Derek

Continuing with my "How to Get Hired" series, tonight I want to talk about getting a job as a new PhD. If you're in this market, the first thing you should do is go back and read that previous post about getting hired with a Master's degree or lower, because most of that applies to you, especially in your first few years in the industry.

That is to say: if you're a chemist, you're going to be making compounds, and if you're a biologist, you're going to be running assays. The difference is that with a Bachelor's or Master's degree at most companies, those things will remain your primary jobs. With a PhD, you'll be doing them and more. (A future post will address the "more" part.)

So, since the starting points for PhD and Master's hires are fairly similar, the same advice applies to your background. As a chemist, you want to demonstrate versatility and problem-solving ability. If you've got plenty of those, we can overlook quite a few other things. Problem is, those aren't the qualities that just leap out at us during a one-day seminar and interview schedule - unless you make them leap, of course, which is your job for the day.

That means that you shouldn't show a bunch of slides of perfect reactions that all worked perfectly every time. Some people do this to show what amazing hands they have, but that's a mistake. No one gets everything to work all the time, not around here, anyway, and you'd better be ready to deal with that. Show us some things that stopped you in your tracks, and show us how you got around them. As long as you didn't get into the initial bind through your own stupidity, these examples will do you a lot of good. Even if you didn't manage to fix something in the end, show us that you took some good cracks at it and were able to move on. You're going to be doing plenty of that here, too.

On another topic, when you're interviewing for a chemistry position, don't try to impress everyone with your knowledge of biology. You're not a biologist, and that's not the position you're being brought in for. If you sound too much like one, we'll wonder if you every got around to learning enough chemistry. We're all convinced - most of the time, with good reason - that academic medicinal chemistry and academic drug discovery don't train you well for what industry is like, so if that's your background, you need to avoid sounding as if you already know all these things it took us years on the job to learn.

But on the other hand, if that's not your background, don't worry about not knowing all the details of enzyme mechanisms, pharmacokinetics, high-throughput screening assays, and the like. Heck, don't worry about not knowing any details of some of that stuff. Most of us didn't know it, either, and we picked it up just fine. You will, too, if you show us that you're the sort of person who can learn new material.

With a doctorate, you'll be expected to show a capacity for independent work as soon as you start picking up the basics of drug discovery, so show us that you're ready for it. It's not going to look good if you got all your ideas by asking one of the post-docs in your group, for example. I don't mean that have to have invented all your reactions, of course. Lifting 'em from the literature is just fine; that's what we spend our time doing here. But did you motivate yourself to go look, or did someone have to poke you?

As I advised earlier, be ready for all the obvious questions about your work. If you show a weird reaction, someone is going to ask you the mechanism, and not only should you know it, you should show that you knew that the question was coming. If you got some unexpected results - and I hope you did - you should have some explanations ready, even if you don't know which one is correct. A real interview-killer is to be asked why you think something odd happened and to answer that gosh, you don't really know, it must have just been one of those things. . .

Comments (4) + TrackBacks (0) | Category: How To Get a Pharma Job

May 9, 2005

Punching the Clock at Merck

Email This Entry

Posted by Derek

Man, is Merck's pick for new CEO ever getting some tepid reviews. I'm agnostic on the subject, but the coverage of Richard Clark's promotion range from resigned sighs to outright handwringing. In that latter category, see this article by Melissa David at

"When Merck crowned a new CEO last week, some disappointed investors saw fresh evidence of an entire industry that has yet to get its priorities straight.

Merck chose an insider without a background in medicine at a time when observers agree that the company desperately needs to focus on developing new drugs. Moreover, Merck made its selection even after its glory faded under a previous nondoctor, Ray Gilmartin.

Richard Clark, tapped last week to replace the embattled Gilmartin, is no master of drug research or even the marketing activities that, to the dismay of some, now seem to drive the industry. He has instead spent his 33 years at Merck concentrating on such areas as manufacturing and information technology."

You know, it's odd, but the chemists and biologists inside most drug companies even see the M.D.s as a little suspect, a bit insulated from the real world of research. Just goes to show what perspective can do. . .mind you, I'd rather have a medical type any day than someone from marketing. (No offense, guys, but we're different species and we both know it.)

Meanwhile, over at Forbes, Scott Gottlieb has a theory about why someone from the production end of the company is on top now: cost cutting. If we can't come up with the big-selling new drugs, then we can squeeze the overhead on what we have. His downer of a wrapup:

"This is one way to view Clark's ascendancy as Merck chief. The company is not preparing for a future filled with its own breakthrough research but, instead, for life as a regulated utility. It is a future where profits are increasingly set by price controls and cutting costs has become the best way to squeeze out extra income."

Well, that's just fine. From a scientist's standpoint, that sounds about as exciting as working the second shift down at the bottling plant. It's going to be up to us in the industry, particularly the R&D end of it, to try to keep this from happening. If we can actually get some new drugs, and whole new categories of drugs, for major unmet medical markets, perhaps we can escape this dystopia. There are a lot of technology bets being placed out there - here's hoping some of them pay off.

Comments (2) + TrackBacks (0) | Category: Business and Markets

May 8, 2005

How To Get Hired: Associate Positions

Email This Entry

Posted by Derek

I thought I'd expand my recent thoughts on getting employed in the drug industry, and start a new category to put them in. Perhaps they'll be a helpful reference to job seekers. I'll be writing from a chemocentric point of view, but I'll try to make reference to other areas of the pharmaceutical job market.

Tonight we'll discuss lab associate positions, since there are more of those than anything. This job goes by different names in different companies, but there's one thing about it that doesn't change: it's a non-PhD position. This is where you'll be with a Batchelor's or Master's degree, count on it.

In a med-chem department, you'll report to a PhD lab head, and you'll be at the bench making the bulk of the new compounds for testing. That's the single biggest requirement for this job: you need the hands to make new drug candidates. You'll turn out strings of compounds inside a given series, then hop to another and run that one for a while.

You'll need to switch between making 20 milligrams of a compound and making ten grams of it at times, and you'll need to do it with some reasonable speed. We don't care as much about reaction yields in the drug discovery labs. Someone who can hack out ten analogs in 40% yield apiece is a lot more valuable than someone who takes the same amount of time to make one of them in 90%.

If you're that kind of person, you'd be happier in the process and scale-up side of the business, where we try to find the best and cheapest way to make the compounds we're really interested in. That's a vital area, and anyone who tries to sell it short to you is not to be trusted. Keep in mind, though, that those folks work under some of the least flexible deadlines in the whole chemistry department.

If you're in drug discovery, which compounds you make and how much you get to move the chemistry around all depend on who you report to and how good you are. Personally, if I have someone competent reporting to me, I take a hands-off attitude, as much as possible - I try to do only the minimum of "Make this, make that" requests. That's clearly not possible with someone just learning the business, though, and anyone coming in right out of school can expect a lot of pretty direct assignments. As you show what you can do, you'll be given more room on your own, and if you aren't, you should look elsewhere, either inside your company or beyond.

That brings up another key thing about these jobs. Let me speak frankly: at almost every company, there will be a ceiling over your head. The PhD is the terminal degree in the field, and you will never rise as high (synonymous with "earn as much money") as those who have attained it. This will be true even if you're smarter than many of them, and even if you could do the job if they'd just let you. In many organizations, for example, you may not ever have to chance to have someone report to you. If that's important to you, you need to choose carefully, or you need to hang in there for the PhD itself.

That's not to say that an associate position can't be a good career. It can be, and I've known people who've done well in these roles for decades, both professionally and financially. But you really should know the score going in.

How do you get hired into this kind of position? Most companies have a mix of Batchelor's and Master's people filling these roles. Naturally, we expect more from the latter, and you should be prepared to meet that expectation. You'll need to show that you've done a variety of different reactions, because you're sure going to be seeing a variety of them if you're hired. It's going to be a tough sell if all you've done are eighty-seven different aldols. For the same reason, it helps to show that you can be a quick study. If you'd never made any whateverazoles or done the Whatsis reaction, but had to pick it up on short notice, be very sure to mention that in your interview seminar. (For more on that seminar see this post.)

And, unfair though it might be, you should be able to demonstrate that you can get things to work. Now, not all projects work, true. But even if you were stuck on a loser project, you need to show that you cranked away at it, tried new approaches when you got bogged down, and succeeded wherever you could. Because, let's face it, there are truly people out there who can't even boil an egg, and we sure don't want to hire any of them. (Or any more of them.) Put the best face on things that you can, without crossing the line into overt bullshitting, because that'll be spotted instantly. You may well be talking with some people who've been doing chemistry for as long as you've been alive; trying to snow-job those folks is not recommended.

You should be reasonably competent in the majority language of the company's labs. It's true that every department has some non-native speakers that are rather hard to make out, but that's not a category that we like to add to. While chemistry can be conducted fairly well through drawings and sign language, if someone runs into the lab and yells "Run for your life!", you'd better not just give them a blank look.

And there's another thing that it may be too late to do anything about: you need to look like the sort of person that people can stand to be around. We don't have one-man isolation labs; you're going to be rubbing elbows with people all day long and working on large teams. If you give off odd or nasty vibrations - and let's face it, some of us in the sciences do - then it'll be harder to get hired. There is, at least as of this writing, no affimative action program for the weird.

Comments (11) + TrackBacks (0) | Category: How To Get a Pharma Job

May 6, 2005

'Tis the Season

Email This Entry

Posted by Derek

Just a brief note this morning. Starting late next week we have that fiesta of spin known as the American Society of Clinical Oncology meeting. A good preview article can be found here. Reading that one, see if you can imagine making investment decisions based on the abstracts of a meeting. How long ago do you think those were written, you itchy investors, and just how much detail do you think anyone is going to include in them?

Comments (0) + TrackBacks (0) | Category: Cancer

May 4, 2005

Another Law of the Lab

Email This Entry

Posted by Derek

This time, it's the libelous assertion that:

You should only believe yields in Tetrahedron Letters papers if you also send off for everything you see advertised on late-night TV.

Well, that's a little unfair. Not completely unfair - just a little. There are papers in Tet Lett whose procedures are perfectly reproducible - I've used some of them. And on the other hand, there are impressive full papers in JACS that have steps whose efficiency could only be reproduced by angels or advanced space aliens. I shouldn't be so hard on one particular journal. But I'll stand by the principle behind this one, and extend it to include other brief communications in the journals that specialize in them. Chemical Communications is a worthy example from England, and Chemistry Letters gets the nod from Japan, and how.

Why the scepticism? I know that the editors of Tet Lett have been trying to remedy the situation, and things have certainly improved from the days when I wrote that law back in grad school. But people often use these journals as dumping grounds of one sort or another. When there's not enough room to write out full experimentals, who can say you're wrong? Three carbon-carbon bonds formed in the same reaction. . .hmmm. . .how about 85% yield? Do I hear 96%? Sold! If the scheme in the paper just has a lithium reagent drawn over an arrow, who's to say that they didn't optimize it out the wazoo with some esoteric blend of solvents and temperatures? Without a full experimental section, we'll never know.

There's always going to be a residue of doubt around a paper that lacks full details. With apologies to the non-chemists in the audience, it's time for some lingo: You say you took off a trityl group and your THP stayed on? Show me. Tell me just how you did that, so I can see if I believe it. You say you got a 94% conversion with that exotic chiral zinc reagent? Peachy! Tell me how you made it - and that includes what kind of zinc it was, and where you bought the darn stuff. I'd like to do that reaction, too, and seeing a bunch of arrows and yields isn't going to help me much.

The idea of the shorter-length journals (or should I say, the ideal) is that they'd be used for preliminary communications of work that would be reported in full later on. Sometimes they are, but I'd like for someone to go to the trouble of seeing just how often that really happens. No one, as far as I know, has ever done that (and I'm not holding my breath, because it'd be a bibliographic nightmare,) but it would be interesting.

Comments (3) + TrackBacks (0) | Category: Lowe's Laws of the Lab

May 3, 2005

Ghostly Influences

Email This Entry

Posted by Derek

Via Matthew Herper at Forbes, here's a real grit-your-teeth article on the ghostwriting of journal articles from inside my industry. Now, I know that this stuff has been going on for a long time in the medical world, and I know that it happens constantly with newspaper op-ed pieces. It's a growing trend/problem in the blog world, too, for that matter.

But all that doesn't mean that I have to approve of it, and I don't. This practice is not only wrong on the face of it, it's counterproductive. (I don't expect to convince anyone who needs convincing just by pointing out the ethical problems, you know.) The medical and scientific journals don't need any more junk in them than they have already, thanks very much. And the pharma industry doesn't need any more opportunities to be seen as a bunch of shady influence peddlers, either. We're already long that position pretty thoroughly, wouldn't you think?

Comments (5) + TrackBacks (0) | Category: The Dark Side

May 2, 2005

Merck and Integra Go the Distance

Email This Entry

Posted by Derek

As many readers will have recognized, I was speaking yesterday about the suit between Merck KGaA and Integra. This one has been going on for nearly ten years now, and is a bit of a mess, but here comes the short version:

In the early 1980s, researchers at the Burnham Institute in La Jolla discovered the so-called "RGD peptides", named for the single-letter amino acid sequences they share. These bind to cell-surface receptors called integrins, which are important in a variety of processes that involve cell adhesion. Blood clotting, wound healing, bone growth, and the invasiveness of some cancer cells all have integrin components. The Burnham group formed a company, Telios, and licensed their patents to it. (Later on, Integra bought the rights from Telios, which is where they enter the picture.)

Merck (Darmstadt) and several other companies also became interested in integrins. In the mid-1980s, Merck collaborated with a group at Scripps (also, as fate would have it, in La Jolla), who showed the potential of blocking some integrin subtypes as a method for inhibiting angiogenesis in tumors. (Yep, anti-angiogenic therapies have been in the works for that long.)

Burnham sued, claiming that use of three key RGD peptides by Merck and Scripps violated their patents. (Licensing discussions had fallen though.) Merck fought back by claiming that their work came under a "safe harbor" provision of patent law, by which companies can do research needed for FDA approval of a drug even if that work actually infringes other patents. A District Court jury awarded Integra $5 million in 2000, later reduced, and Merck appealed. The Federal Circuit Court didn't buy Merck's reasoning, either, and ruled that the safe harbor language wasn't meant to extend that far back into early drug discovery and lead identification. (A longer discussion of that ruling can be found here.)

Merck appealed again, and the Supreme Court agreed to hear the case. That's unusual; most patent disuputes don't make it past the Federal Circuit. But the court seems to think that the safe harbor language is both ill-defined and important enough to deserve their attention, and I think that they're right. Merck (and several other drug firms who've filed friend-of-the-court briefs) argue that if the safe harbor provision is defined that narrowly, then big swaths of drug research are going to have to either come to a halt while patents are hashed out, or it'll just up and move to some country that doesn't respect the IP.

Meanwhile, Integra (and other companies who've lined up with them) say that if the law were interpreted according to Merck, then there would be no incentive to companies to search for and patent new research tools. After all, if you could claim the FDA exemption for anything that might end up leading to a drug - that is, anything that might lead to someone making some money - then what's the point?

Oral arguments took place on April 20th. I'm going to bet that the decision will split hairs in such a way as to not shake things up too drastically (as the court did in the Festo patent law case), but I'm glad I'm not betting with the kind of money that Merck and Integra are.

Comments (9) + TrackBacks (0) | Category: Patents and IP

May 1, 2005

Between Friends. Sort of.

Email This Entry

Posted by Derek

There's a patent issue that I've been meaning to write about for a week or so. All I have time to do tonight is set the stage for it.

Let's say that Company A has an interesting compound, a clinical candidate that's moving along and showing interesting activity. It's out there in the patent literature, and they've spoken about it enough to release its structure. Naturally, the competition is interested. Can Company B make themselves a batch of the compound to see how good it really is?

Well, I hope that the answer remains "yes", because we do that sort of thing all the time. We're always sizing up the competition, and they're doing to same to us. It's not unheard of for scientists with friends at other companies to get calls from them saying "How come you people say that your compound does so-and-so? It sure doesn't for us, you losers!"

What if Company B wants to use Company A's compound to do a little bit more for them, though. . .like, say, serve as a crucial ligand in a screen of their own compound library? And what if Company B uses those results to eventually discovers a fine drug candidate of their own, one which, arguably, they wouldn't have found at all without the use of Company A's chemical matter? Is there a problem here, or not?

Well, the Supreme Court has ended up with a case that bears on just these sorts of questions, so later on this year we can expect to all be enlightened, enraged, or just plain baffled. More on this tomorrow. . .

Comments (5) + TrackBacks (0) | Category: Patents and IP