12-23-2016, 01:46 PM
(12-19-2016, 10:29 PM)Rynn Wrote: In a nutshell the answer is: difficult to extremely-difficult. Most phenotype traits are an emergent consequence of thousands of interacting genes and the environment. To get to the type of reliable, radical (i.e. leads to speciation) genetic engineering we have in OA will require a far greater understanding of developmental biology and phenomics. When we'll have that knowledge IRL and what consequences there will be is anyone's guess.
That said there are many things that would be simple and we are technically capable of doing them now. There are plenty of traits that are determined by a small group of genes or even a single gene. Combined with genetic epidemiology (a field that's going to explode the cheaper DNA sequencing gets) we could find strong correlations between certain gene combinations and traits and be able to engineer them even if we're not entirely sure what the mechanism behind it is.
This reminds me of the idea of a 'geneweaver' that I was kicking around some years back.
Basically we presume that at some point fairly early in the setting, Terragens made a number of breakthroughs in understanding how all those thousands of genes interact and developed programs that could model the process with a very high degree of reliability. The resulting tool is a device that you input your desired traits or end result lifeform into and it figures out the combination of genes (natural and perhaps eventually artificial) that are needed to produce it. It also models the resultant lifeform virtually through some huge number of iterations and makes adjustments as needed if these turn up any issues or long term problems or the like.
I should see about picking that article up again at some point.
Todd