Biology's Master Programmers

Human-engineered circuits are easy to predict; they were designed in a way that made the most sense. With biology however, something that appears to be working may have a completely unrelated side effect that isn't noticed until decades later, or even something so bizarre that it isn't noticed until hundreds of years later in future generations.

The way we currently find treatments (grossly oversimplifying) is to grab some substance that looks like it might work, test it in animals, test it in humans, wait a while to make sure nothing really bad happened, and then prescribe it as a drug.

If computers get advanced enough, I would much prefer the case where all of these interactions are fully simulated. You can go through millions of variations of substances all while tracking every change they make to each part of the virtual organism. Obviously, this is a VERY long way off, but it would be nice because we wouldn't have to worry about unknown side effects.

Of course, our current way is the best we've got, and it saves millions of lives each year, so while there may be some future risk, we're doing the best we can.

*On a somewhat related note, I've always thought it would be cool to have programmable blood filters. In other words, you could essentially provide the type of molecule or cell you are looking to trap, and then something like a ccd would either redirect it or let it pass through.

Leo Laporte did a TWiT Special on this recently: http://twit.tv/show/twit-live-specials/126

The discussion centered around taking existing genomes and splicing in functionality. The example given was setting up a Kickstarter project for trees that glow in the dark. The government is aggressively pursuing this tech.

[Edit] Direct links:

dna editor: http://www.genomecompiler.com/

dna printer: http://angel.co/cambrian-genomics

Once you have a scalable process for producing chemicals that are potential drugs via DNA sequencing & analysis and bacterial synthesis, how do you figure out what disease process to test those chemicals against? Do you shotgun them against a large set of disease processes? Purify the product, analyze it, and store it in a database and do theoretical work to pick out candidate diseases for the chemical? Is there something about the gene cluster that gives clues as to what the chemical might be doing for the organism?

Sounds like Intelligent Design.