What sort of chemistry should people be spending their time doing in drug discovery? The obvious answer is, “Whatever makes a drug!”

However, if you think that this is sufficient, you may have been spending too much time in management seminars or listening to expensive consultants. If we knew what made a drug, we’d be doing it already, and a bit more often than we do now. As has been all too clear over the years, we don’t quite know which direction we’ll end up going when we start a drug discovery project, so one has to be ready for any sort of chemistry.

For better or worse, though, any sort of chemistry often seems to mean plenty of palladium couplings and plenty of amide formations. I plead guilty myself. The last half-dozen reactions I set up in my own lab fall into these two categories. In my defense, these really do make sense for exploring the lead structures I have on hand, but everyone else probably thinks the same way, too. Given the sorts of structures that come out of a lot of screening campaigns, all of us are probably right, for the most part. The hits from the deck were likely made via yet more palladium couplings and amide formations themselves, so that’s the natural way to make analogs around them.

But there’s more to life than these reactions.

“That’s for sure,” some readers are thinking, “there are sulfonamides, too, and even reductive aminations, for when you’re feeling adventurous!”

Fair enough, although speaking personally, I’ve mostly found sulfonamides to be nothing but trouble in the end and yes, I know that they’re out there in marketed drugs. Still, there’s a lot of chemical space that can only be explored when we get beyond these and the other layup, stand on one leg, guaranteed to work transformations. It’s always a judgment call, because you can cover so much more ground with the easy reactions. What you have to make sure of, though, is that it’s ground that you really should be covering, or if you’re just heading that way because it’s so darn convenient. At that level, this is not a problem that’s specific to organic chemistry, of course, but it applies just as well as it does to other optimization problems. The road has ditches on either side of it—you can go off into one of them by ignoring the boring, pedestrian stuff on principle, even when it’s what needs to be done. And you go off into the other by doing nothing but the easy parts, while congratulating yourself that you’re making so many compounds and doing so much work.

Convenience can save you time and money, but it can also cost you. There have been several papers over the last few years that show an apparent relationship between the types of compounds you get by lots of metal-catalyzed couplings—that is, lots of aromatic rings—and downstream trouble with PK or toxicity. I’m not as sold on this as I once was. The correlations vary depending on whose data set you use, and toxicity is such a complicated subject that it would be a bit surprising if it were subject to such a simple rule. But well before that stage, you can certainly get into trouble with the in vitro assays by making a bunch of insoluble compounds, and palladium couplings can certainly deliver those. There aren’t many assay technologies that can’t be messed with by a sample of hazy, fluocculant test sample, and when that snowglobe effect is happening in straight DMSO stock, you’re in trouble. There’s also an inbreeding problem that develops in a lot of company screening collections, alluded to above. If most of the screening deck was made by the same sorts of chemistry, most of your hits are going to have some broad similarities as well, which may not be a good thing.

All these convenient reactions have another effect as well. Some of these are getting so reliable and easy to perform that they’re becoming a subject for automation. This has been coming for some years, with a lot of dead ends and false starts, but just because it’s taking a while doesn’t mean it isn’t still on the way. The bread-and-butter reactions of organic chemistry are getting into the area where humans don’t necessarily need to perform them, especially when they’re being done in something like the exploratory phase of a drug project. Outsourcing was the initial shot across the bow in this area: you can get your amides and sulfonamides made a lot more cheaply somewhere else, in many cases, and the machines will be even cheaper. This same progression has happened in many other kinds of labor that once needed human hands at every step, and there’s no reason why it can’t happen to at least parts of organic synthesis.

What this means is that medicinal chemists should be spending their time, as much as possible, on more high-value activities. Most drug research takes place in high-wage countries, and everyone involved eventually has to justify those costs. In terms of organic synthesis, that could mean mixing in some harder reactions as a break from the routine stuff. If that sounds too worrisome as a hit to productivity, then perhaps there’s another way: looking for the “new easy” reactions. Staying current with the literature can help a person find transformations that don’t get used much in the drug labs, but are useful and fairly general as well. We medicinal chemists should be pulling in new methods from the academic labs as often as we can, because reliable ones can take a project into chemical SAR space that might not otherwise get looked at. They can also take your project into new IP space pretty quickly, which is not an advantage to be taken lightly.

On a more personal level, this sort of thing is better for a person’s career. You’d want, ideally, to be able to tell a prospective new employer about being current with the field, ready to try new things, and able to get them to work, and this is a good way to do it. Drug discovery has a lot of potential ruts that people can get stuck in, and while conditions in the industry are not being particularly kind to anyone, they’re especially hard on people whose main selling point is that they can provide what a lot of other people are already doing.

I’ve been talking about organic and medicinal chemistry so far, because that’s my own field. But the same considerations apply in other departments as well. New methods, new assays, new instrumentation, new concepts—all of these should be as large a part of your job as you can reasonably make them. If your current employer doesn’t seem to value this sort of thing, you might want to make yourself look worthwhile to someone else who does. It’s not only just good for the brain!