Not that long ago, my laboratory was doing its usual thing, cranking out the potential Wonder Drugs and sending them downstairs for testing. The results come back in a batch, and we all check to see how everyone's compounds fared: "Hey, how come that one's good when the other one isn't?" "Man, I'm glad I didn't make that one down there. . ."

One of the folks in my lab had a series of reactions to make her compounds that had one tricky step in it - the group that was being added on could potentially end up on two or three different spots. You'd think that would be reason enough to avoid that route, and generally, it would be. But plenty of other compounds had been made that way on the project, and everything had been fine - only one isomer was produced, for some reason. We'd become pretty complacent about that step, actually. But this time, we were doing it on a somewhat different core structure - and yep, it was different all the way through.

This one made a mixture of two compounds in the reaction, one more than the other, and figuring out which was which took some doing. After staring at it for a while, we turned the problem over to our NMR specialists. They took the thing apart with all sorts of neat spectroscopic routines and came back with an answer: we'd made two of those potential isomers this time, as we'd figured, but the one we'd been shooting for was the minor one.

Well, that's chemistry as my colleague and I are used to it. And since both of us have been around the industry a few years, we decided that she should just go on with both of them. She elaborated the molecules through to the last step, and then we sent them downstairs to be assayed for activity. What the hey - she'd made the stuff anyway, and we might as well get some use out of it. (Of course, we had no idea what the target protein would think of this rearranged layout in the molecule.)

When we got the results back a few days later, it turned out (you saw this coming) that the new "wrong" structure was about five times more active than the "right" isomer. Oh, we know what we're doing, all right, we steely-eyed drug designers. Never believe it's not so.