GeneSmith

Comments

The Case for Human Genetic Engineering

I'm thinking of something like a fitness trap scenario, where competition to maximize zero sum traits degrades some other key trait in an irreversible way. Not that it would literally be irreversible, but that the degradation of such a trait (perhaps we find a gene that makes you very attractive but dumber) would make the next generation even more likely to sacrifice that key trait etc etc in a vicious cycle.

I'm thinking here of the Irish Elk, a huge species of deer whose competition for larger antler size drove it to extinction.

See here: https://www.nationalgeographic.com/science/phenomena/2008/09/03/the-allure-of-big-antlers/

Though I agree with you that the danger of banning genetic modification would be much, much greater than the danger of this kind of sexual selection induced extinction.

EDIT: After reading the article I linked it looks like there is actually controversy about whether large antlers drove the Irish Elk extinct. The real cause may have been a combination of a reduction in food an predation. So perhaps that's not the best example for the wisdom of banning zero sum trait selection.

The Case for Human Genetic Engineering

I've spent a fair bit of time thinking about the potential implications of a soft or hard ban on these types of zero sum traits. You're probably right that people wouldn't accept mandated downgrades from their current possession of these zero sum traits (shorter, smaller breasts etc), but it seems plausible that at some point we might put a cap on how extreme we're willing to let people engineer themselves.

But historical precident has given me pause. One can imagine that the gigantic benefits to the species as a whole of increased intelligence would not at all have been apparent for most of human history. Might we accidentally ban a trait that appears to be zero sum but actually has massive positive externalities that we simply don't foresee? That's one of the things I'm worried might happen with these types of bans.

Of course there are probably even bigger risks if we simply allow unlimited engineering of these sorts of zero sum traits by parents thinking only of their own children's success. Everyone would end up losing.

The Case for Human Genetic Engineering

I am quite optimistic about our ability to increase both intelligence and longevity for a few reasons:

1. Intelligence correlates quite strongly with longevity. It therefore seems likely unlikely that optimizing for increased intelligence will have a negative effect on lifespan in the short term.

2. Increasing healthy lifespan will probably lead to smarter average citizens even if there's no direct effect on IQ. This is simply because a lower percentage of citizens will be in the development phase.

3. There's a paper I read a few weeks ago suggesting that there's far less pleiotropy (genes that have two distinct effects) than we previously thought. If this is the case, we would expect many genes to affect mostly longevity rather than nearly all affecting both longevity and intelligence.

As far as the "this generation vs next-generation" question, I think that there ARE probably some changes you could make to the current generation, particularly by modifying stem cell populations. But there are always going to be some changes that can only be feasibly made in embryos (particularly those that affect developmental growth). Until we get extremely advanced nanotech, those types of changes are only going to have an effect if made in embryos or young children. And if your nanotech is that advanced this all may be a moot question because you might just synthesize entirely new designs for bodies and brains that aren't even possible by simply modifying genes.

There's also a knowledge constraint on our ability to modify the genes of adult humans. For highly polygenic traits in particular, we know the rough regions that correlate to the expression of the trait, but in most cases, we don't know the precise gene in that region that is responsible for the observed variance. We'll have narrowed it down to 10 or 11 candidate genes, but we don't know the single gene responsible.

This makes genetic engineering of such traits highly impractical with tools like CRISPR because we'd have to cut and replace ten times the number of genes that we actually want to replace. And given CRISPRs continued issues with off-target edits, this is likely impossible with current technology.

It's much easier to modify such polygenic traits with embryo selection methods, which as the name suggests, can only be performed by selecting one embryo out of a large batch.

The Case for Human Genetic Engineering

Thank you for pointing this out. I was aware of genetic drift, but I hadn't read before that it accounted for the large majority of mutations.

I think what I was saying still largely holds true though: even if gene B has a neutral effect on reproductive fitness, the lack of fitness ADVANTAGE will mean that any spread that does occur will be by pure chance and is liable to being reversed by the same chance.

I actually don't really know how to do the math on this one. If we start out with a population that all has the normal form of gene B and we the mutant form conveys no net reproductive fitness benefit or downside and the likelihood of each mutating into the other is equal, then I suppose we would expect the frequency of each variant to approach 50% given enough time.

Which makes me think that the likelihood of one allele spontaneously mutating into any other is probably pretty important. In fact I know of a specific disease in which the mutation from one allele to another in not symmetrical: Huntington's disease.

Huntington's disease is a codon repeat disorder, meaning that the mutant gene causing the disease has a codon that's repeated at the end a large number of times. People with the disease have the letters 'CAG' repeated at the end of the gene at least 40 times. The more repetitions, the earlier the effects of the disease begin to show. There's actually two villages in Venezuela (Barranquitas and Lagunetas) where children as young as ten acquire the disease due to having 70 or 80 CAG repeats.