Intelligent Design and Common Descent

[Dan_Eastwood]

Does it make sense to think that de novo genes should be more common in the past? Consider a time when these genes did not exist, the potential survival advantage may have been greater than it is today.

[Mercer (John Mercer)]

@Dan_Eastwood

Does it make sense to think that de novo genes should be more common in the past? Consider a time when these genes did not exist, the potential survival advantage may have been greater than it is today.

That does make sense. In addition to your consideration, we should consider that any new function that confers a selective advantage doesn’t have to work well at all. That’s why the size of any functional “islands” in sequence space is important to understand, as they would be landing sites for the unoptimized new functions of new genes.

[swamidass (S. Joshua Swamidass)]

@Dan_Eastwood one of the challenges of studying de novo genes is ascertainment bias. For purely technical reasons it is very difficult to pick out de novo protein when they arise. We can’t use homology to help us detect them, and have to rely on difficult and unscalable direct detection methods instead.

[Dan_Eastwood]

@Mercer

we should consider that any new function that confers a selective advantage doesn’t have to work well at all. That’s why the size of any functional “islands” in sequence space is important to understand, as they would be landing sites for the unoptimized new functions of new genes.

If the function is truely new, there is no competition. Any level of function provides (potentially) a new niche to explore. In a mathematical sense, a new function opens up a new way to explore functional space, and other functions which were once “distant” (mutational distance?) may now be easier to reach.

@swamidass

one off the challenges of studying de novo genes is ascertainment bias.

I was trying to think about this from the mathematical side, and wondering what is means to have “no function”. If a sequence really doesn’t nothing at all, there is still a biological cost to having it, giving it a disadvantageous function (energy cost with no benefit).
Aside from that, even filling empty space might be a function if it changes the timing of some other reaction. It’s not obvious to me that any protein could really be neutral to all function - it’s just a matter of what function it might accomplish.

[Mercer (John Mercer)]

@Dan_Eastwood

If a sequence really doesn’t nothing at all, there is still a biological cost to having it, giving it a disadvantageous function (energy cost with no benefit).

Not if a greater cost is required to eliminate it. Evolution is very inefficient that way.

[Dan_Eastwood]

@Mercer

Not if a greater cost is required to eliminate it. Evolution is very inefficient that way.

Interesting - In my noodling around with genetic algorithms I’ve noticed that rules for how the population is culled can affect how the optimization advances. Keep extra information in the population isn’t necessarily a bad thing.
I need to work that into the Complexity Analysis topic (an ongoing project).