Kaichos Man writes:
So what is the absolute minimum novel genetic structure required to be "seen" by natural selection and added as new information to the genome?
Who cares? It is irrelevant, the genome does not give a rat's ass how "complex" it is. Mutations that reduce complexity and increase survivability are going to be preserved through natural selection just as much as those that increase complexity and increase survivability.
"Complexity" can be preserved through natural selection for completely different reasons. For instance, suppose there is a mutation that causes a long string of random, non-functional DNA to be inserted harmlessly into the genome. This mutation happens to take place in an organism that has a completely different mutation that is simply a modification of existing DNA (no change in "complexity"), but which makes the organism much more likely to survive and reproduce. Behold, natural selection is preserving a lot of completely non-functional DNA, yet increasing "complexity" by your measurement.
That example could have a later organism that deletes the useless DNA, and also has another different survival advantage modification. That would be a decrease in "complexity" and yet natural selection still preserves it.
The question is pointless, nearly meaningless, and completely irrelevant to the issue.