I actually wasn't at work today - took the day off so that we could take the kids to the zoo. That probably means that something important happened; these things always seem to take place when I'm out of town.

And the rest of the time something important happens in the lab, I generally don't even realize it. It's not because I'm especially dense (no superfluous comments, those of you who read this site from my company!) It's that we usually only recognize important scientific moments in hindsight.

For instance, I've never been on a project when we've made a compound, tested it, and said "That's it. That's our clinical candidate right there." For one thing, the tests that a clinical compound has to pass can be spread out over several months. You need several weeks just for your minimal toxicity testing (and holding your breath that whole time, which is what you feel like doing if you have any sense, is a real strain.)

Another reason is that we always figure that we can do better. Every good compound seems to open up a whole new area to work in, and everyone jumps in and adds their two cents to the structure. It's only after you've run through those variations and tested them that you realize that the compound from a couple of months back was as good as things were going to get.

That's why most compounds sort of seem to back into the clinic. They're sometimes just the best of a bad lot, but they're always a retrospective pick, made by balancing the various factors and taking the least obnoxious candidate. A is going to be the easiest to formulate for clinical trials, but B is a lot faster to make. But C is almost as good, and comes from a cheaper starting material. Meanwhile, D had a better tox profile than the others, but was the difference real, or just noise? And so on. . .

It's the same principle as in engineering: you can't make it perfect, just as good as you can in the time and with the tools you've got. And if you don't get the thing out the door, it doesn't matter how good the data might have been. This isn't a license to do sloppy work - that's an even worse mistake. What it is is the lack of a license to go on refining things forever, long after they've reached the point of "good enough."