 So a typical company here knows exactly what a target is that we know a ton about the biology of the target And I also have a way to identify is this target working or not What you then typically do is that you develop something so that we can have a machine a robot test this on small plates So you have these small plates with maybe 512 or even a 2048 or so small wells and in each of those wells I might have one copy of a cell or something and then I'm measuring for instance How much is my iron channel conducting or do I see fluorescence of this particular molecule if it's binding a molecule? So a so-called assay a simple way to test this thing was working or not Am I binding or ideally do I have the right biological effect? And then I just start testing So if you're a large pharmaceutical company, you typically have databases of many many many drugs Maybe hundred thousand two hundred thousand even half a million drugs So these are just small relatively standard compounds that we have produced every drug You produce cost a bit the first time you use it, but it's not that expensive to create more of it And the idea is that if I now have a new target I can take this drug and I can go in I might not test it against my entire database But let's say that I tested against three hundred thousand compounds or so and you know if I'm lucky I might even get a hit So maybe this is an actual example that a colleague of mine was involved in a few years ago This is a compound called lactamase the screen size the number of compounds we tested here was three hundred thousand Unfortunately, I didn't get any hits. Ah, too bad. Can try it again No At this point your boss is not going to be overly happy Because we are talking about actual chemical tests by a robot here It's even if even if we own these chemicals in-house every test here probably costed the ballpark of one dollar So in this actual study people wasted three hundred thousand dollars on one test where we didn't get a single hit We don't have a clue Literally wasted three hundred thousand dollars oops There was another molecule Curtain Screening as a slightly smaller database in this case they found hundred forty hits So here the scientists are really happy But these just these two studies is half a million dollars probably run over a couple of days There is there are two important things earlier first We're not guaranteed to see any hits you can't afford to test everything There are probably databases here that are five million compounds or so But you can't afford to spend five million dollars once a week because you would be fired from the company So we have to try to be wise and just hope that we can have a small set and maybe find something The something is the other keyword. I don't need to find a super plus plus good hit here Sure if I do I'm gonna be happy, but all I care about is finding something. You know these hundred forty six compounds That's something to work with I'm gonna look at them break them up on screen and see what are the common factors Based on that. Can I find other molecules like those? The only part I don't want to end up with is the first line here because if I don't have a clue I don't have a clue so Finding hits is easier than you think It can be harder than we think too, but we don't need something great here We just need something to work with and hundred forty six is certainly something to work with But what we do for lack dummies where we don't have anything to work with we can't keep testing more and more So then we might have to use other techniques in particular computers