A process of synthesizing or "designing" drugs (i.e., pharmaceuticals) out of knowledge of the structure and function of proteins. Pioneered at MIT, UCSF, and the hundreds of biotechnology companies that have spun out of them.

Rational drug design can only be understood in the context of the traditional drug discovery process used by large pharmaceutical companies, a.k.a. Big Pharma. Typically, you send lots of people out to areas of the world that have lots of biodiversity, like tropical rainforests, since Nature is much better at synthesizing new organic material than you are. They take samples of the local flora, scrape mold off rocks and trees, bits of soil, etc. These are then taken back to the laboratory where thousands of bench scientists test them for reactivity with various drug targets, ie: HIV, malaria protease, etc. Promising candidates are refined, the active ingredient is identified, and a drug is born. The downside to this is that it is essentially random, and although there is a lot of really hard science in the testing, isolation and extraction, you are essentially hoping to trip over something interesting. Most drugs are still developed this way.

With the rise of biotechnology and fast computers it became possible to approach the problem in a different way: try to build a molecule from scratch that can attack your drug target. This is on the face of things unbelievably difficult, because typically the three-dimensional structure of the target protein probably hasn't been solved (this is done by crystallizing the protein and using X-ray crystallography techniques), and since structure equals function for proteins, you need to know at least something about the protein structure. Typically you compare the linear structure of the protein to a database of solved structures and come up with an approximation.

Now you take what you know about the 3D structure and look for binding sites, which are usually pockets or other areas where small molecules can bind to the protein to disable it. You then compare that to a huge database of small molecules and look for probable candidates that will "fit", taking into account both physical space and the physics involved in binding. This is a bitch computationally, and is an area of intense research.

A few thousand best fits from the above are chosen. These candidates are then rendered on a graphics workstation along with the target, and scientists settle on the best candidates based on intuition and experience. These are synthesized, sent to the lab, and tested on the compound. The best few of these are given to synthetic chemists who create small variations, and these are tested.

With the knowledge of what worked and what didn't, the 3D model of the active site on the target can often be refined, and another iteration of the process is started. The idea is that eventually the process converges on a drug.

One of the advantages of RDD is that you can limit your attention to small molecules, which are orally bioavailable, ie: you can take them in pill form. Above a certain size, a molecule can't be absorbed into the bloodstream through the stomach, and has to be injected, which is a drag, requires a medical professional, and limits the marketability of the drug.

Log in or register to write something here or to contact authors.