Never send in a chimp when a human will do: assays in discovery research

What a day! Between the scintillating launch of the new blogs (really, I am all aquiver) and doing my best to be irksome in my actual day job, I am more than ready to knock back a dry vodka (Grey Goose, preferably) martini at my favored watering hole in Einsteinville.

Part of my job is to pass judgment on protein targets gearing up for screening campaigns. "Screening" refers to high throughput screening which is the bread and butter of discovery research in Big Bad Pharma. It's an automated process in which the general idea is to increase that needle in the haystack factor. This is accomplished by assaying huge numbers of compounds (typically 1 million plus) against an enzyme or receptor target of therapeutic interest.

Just as the saying goes in techno lingo, garbage in, garbage out. To obtain hits (those compounds which affect the enzyme's or receptor's activity) which follow through nicely in secondary analyses, the assay must be good, and not merely adequate. A foible of the local corporate culture I have encountered here is the rush to get an assay in place for screening so that deadlines in a given therapeutic area can be met.

Well, good science can't be rushed in such a slovenly manner. If it's to be performed with efficiency and accuracy, then the folks involved need to know what they're doing. And there's the rub. Everyone and his/her freakin' cat believe they can develop robust assays. Thus, we have biologists who are otherwise quite talented with things cellular, pharmacological and organismal, who really don't have a good grasp of assay design. They don't check out linearity of assay response with time, protein concentration, influence of buffer components, the identity of small molecule and protein co-factors, etc. At some point along the line, they discover something isn't right, and my group, the biochemists (we focus on mechanistic studies), are called in for consultation. Ideally, we'd be there in all the exploratory trenches setting up the assays ourselves, but we're a small group and well, we just can't support everything and everyone. Therefore, we are sometimes doomed to consult. Shudder. So here I was consulting this afternoon.

The issue was an enzyme which, in my much vaunted opinion, is FUBAR . Its specific activity is sucktacular at best, and the Km of a key substrate is so far off the mark as to be in orbit around Jupiter. Yet, the therapeutic area wants to get this into screening ASAP, and one of the guys in the screening group is hesitant to tell them "No f*cking way." My opinion was solicited. So it fell to me, biobitch extraordinaire, to carefully word an e-mail which, in the nicest possible terms, said "Your enzyme is crap. Time for a do over!" Additionally, since this enzyme will be brought into my department for detailed mechanistic analysis, I'd really like to have some good protein in hand.

I hit "send."

Mine will be an unpopular judgment to be sure so we'll see what falls out next week. However, the chemists are very concerned about spurious data emanating from this particluar protein. And I always want to keep the medicinal chemists happy even if I have to piss 'em off sometimes.

All righty then. I'm off for my martini and to contemplate the words of a former mentor (a rather brusque and opinionated fellow):

Any monkey with a cuvette and a Gilford can "do" enzyme kinetics. It takes a human to interpret the data.