Let me caveat this first by saying I understand that eugenics is utterly reprehensible, morally and ethically.
But is it practically/scientifically viable? If so, how?
And if so, how long would it take for a significant (advantageous) change in overall human genetic composition to become evident, or would there be unintended consequences?
Some forms are, yes. One example is using CRISPR to edit out genetic diseases and replace those diseases with healthy genome.
Eugenics in the 19th Century model (that is, selective breeding) is possible as well, but it would have to be run by a very stable institution with unchanging long-term goals (that is, bring out these specific traits, repress these specific traits, after demonstrating that those traits are genetic and that it’s possible to accomplish their goals with their seed stock) because any such project would be, by definition, multi-generational. Such a project is not necessarily doomed to fail, and it’s not necessarily doomed to be Literally Nazi.
My point is that, one, you probably don’t consider gene editing to be eugenics, even though it’s the only feasible form of eugenics on any species with a human-scale generation time, and, two, most people assume that the old-fashioned selective breeding form of eugenics must needs be done with old-fashioned unscientific racism as its founding motive. Once you move beyond those ideas, you can see successful eugenics all around: Literally every food crop is the result of eugenics, as is every single dog breed, by definition. Ditto cattle and horses and most of the other animal species we ranch or consider pets. Humans aren’t special, we just live as long as, well, humans, so any selective breeding program centered around changing the allele frequency in a human stock would have to be a long-term affair.
Worked on dogs.
It didn’t make them “better” per se, but it certainly changed them in accordance with the desires of the ones controlling the breeding, over time. “Better” and “worse”, to the extent that we can confirm, are human constructs - effectively, opinion - and vary across time, space, culture, and individual.
It isn’t always wicked. I know several people who have chosen not to have children, because they know they are carriers of unpleasant genetic syndromes. They are, by choice, improving the overall human genome, in a very minor way.
The word got a nasty connotation from people who thought to remove the element of voluntary choice from the equation.
I take it that you’re ignoring practical considerations at the political/social level in the sense that you assume that there is complete control over the breeding habits of the population.
That being so, it would be straightforward to virtually eliminate monogenic diseases in one generation by screening children before puberty, and stopping anyone who carried (say) the cystic fibrosis mutation from breeding. Some maintenance would be required in subsequent generations to screen for de novo mutations. There are rare cases where a “disease” may be prevalent because it’s beneficial under certain circumstances (the famous one is sickle-cell, which may be protective against malaria), but virtually all monogenic diseases are simple loss-of-function mutations, so it’s hard to imagine any unforseen negative effect from eliminating them.
If you wanted to (say) change some trait such as the size of the average human, then you’re in much trickier territory. It could certainly be done, up to a point, but there would likely be unintended consequences in other traits, as we already know from (say) dog breeding. Most traits involve the complex interplay of many genes, and many genes have widespread effects on many traits.
You could certainly breed humans for specific traits. It would be hard to make a generally “better” human though.
You could maybe make us stronger or faster runners, but there would likely be trade-offs. If you tried to breed smarter humans, you might have to abandon natural childbirth as the width of the birth canal seems to be a limiting factor on skull size.
Who knows. They may be the one couple who produces the key combination of unrelate genes that will save all of humanity at some point in the future. That is, there is no way of knowing if they are improving the human genome or not.
Since there is no reason to believe that this magical combination of genes will arise in this couple’s offspring rather than any other, the statistical expectation from their action is simply to remove deleterious mutations from the population. We act based on statistical expectation all the time, without complete certainty. I don’t think this is a good argument against their action, or against eugenics in general. (There are other good arguments against eugenics, of course!)
Here’s a thread from 8 years ago on the same point: What, scientifically, is wrong with eugenics? - Factual Questions - Straight Dope Message Board
The problem with this is that virtually everyone is a carrier for several deleterious alleles. And that means you have to cull a very large majority of your breeding population. And since your breeding population is humans, they aren’t going to stand for that.
And it’s silly and senseless anyway. The point of eliminating the gene for cystic fibrosis is to eliminate cystic fibrosis, yes? And so we test everyone to see who’s a carrier, yes? Then all we have to do is counsel people who are carriers to not have children with other carriers, and if you happen to be married to another carrier and you can’t stand the thought of not having a child with them, go ahead and have the child and we’ll test the embryo and if the embryo is homozygous we abort it.
I mean, we are talking GATTACA level ubiquitous genetic testing here, so if we’re doing that the vast majority of hereditary diseases can be eliminated without culling the carriers of those diseases. The point is to eliminate genetic disease, and we can do that a lot more easily than forcibly sterilizing 80% of the population.
The way eugenics was explained to me so that, for the first time, I could understand why it was a big issue, rather than being distracted by talk of crops and racehorses, was when it was boiled down to a conflict of individual’s right to do what they wanted with their body, including reproductive rights, versus a view that there was a higher mandated social good that constrained individuals’ choices.
In that context CRISPR technology if applied with consent, parental decisions not to have children, choosing your partners so you can breed the next Usain Bolt are probably not eugenic, while one-child policies, terrorising women who wish to terminate pregnancies and forced sterilisation for people with a mental disability are somewhere in the eugenic spectrum. With the latter, perhaps there is a role for the state as the ultimate guardian of those without capacity to make sound judgements to act to minimise overall harm, but I see no legitimate state or social purpose for the others.
Overall though, the question of would it actually work would probably be no, because at least where this forum is read, that contest between collective and individual rights is played out regularly and neither side will win enough times to make a cumulative difference.
I thought the hypothetical was that we were ignoring these considerations. If we ignore social considerations and treat humans hypothetically as though they were lab animals, we certainly have the technology to screen the entire population for the most prevalent monogenic diseases at moderate cost, and that’s much simpler than screening all fetuses with selective abortion. From a genetic perspective there’s no reason that we would not choose to breed from only part of the population.
Discouraging carriers from mating is already being done with regards to screening for Tay Sachs within the Ashkenazi Jewish community through the Dor Yeshorim program.
The catch is that such a screening program may actually increase the prevalence of a deleterious allele in the population. If test results are used to discourage two carriers from having children then over time more and more people are carriers. The test eliminates natural removal of alleles from the gene pool experienced when a person with two detrimental alleles dies without having children.
This could partly be ameliorated by using advanced reproductive techniques to select against both heterozygous carriers and those embryos which are homozygous for the deleterious recessive trait.
I don’t think this is right. You’re correct that doing nothing does eliminate any child with two detrimental alleles from having children. But such a child only comes form two carrier parents. If you have those two parents not have children, you eliminate all their two detrimental allele children before they’re born so obviously they won’t have children, and it also eliminates any of their children who would be carriers. That is it eliminates all future offspring that would have copies of either of those detrimental alleles.
I might make things worse, if it discourages those to people from marrying each other so they each marry someone else and continue to pass on their bad gene and don’t eliminate any two-detrimental-allele children, but ti might also discourage either one from having any children at all as well.
There’s a more efficient way to do this.
Get a genetic database of all major genes across thousands of people who you know reliable information about. (how long they lived, every illness they ever had, their IQ, success, physical stats, etc)
First, correlate the same gene with all the other examples of it. Analyze like following :
If a gene has 3 variant alleles with a frequency of 60%, 40%, and 0.1%, you can safely eliminate the 0.1% allele right off the bat. Almost certainly (statistically this is correct), the 0.1% minority allele is a mutation that is not helpful. If it were a beneficial mutation, it wouldn’t be 0.1%, it’s most likely a mistake that is either neutral or deleterious. (the odds of it being beneficial are very low, so this is a safe bet)
Now, for things like 60/40 splits, look for commonalities across the population to try to figure out what the gene does. In some cases, there won’t be anything you can measure that has statistical validity, if so, just leave it alone.
For other genes, we know what they do from other studies. For example, certain immune marker genes we actually know how they work, and we know that variations are good. Giving each ‘designer baby’ a unique MHC combo that we know is valid because it isn’t correlated with problems like allergies would make sure our uber-babies aren’t vulnerable to disease.
We can readily guess a gene’s function if it translates to a known peptide chain or is a promoter for one - so if it codes for a different variant of myelin, that might be a good thing.
So after you do this first cleanup pass, eliminating errors and picking the better version of each allele when the benefits are measurable, you start making crops of clone designer babies. You make them in batches - you want 100 or 1000 or so genetically almost identical babies for the next phase. The almost is you do shuffle those immune system genes so a disease can’t kill them all.
Now, in the next phase, these uber-baby clones should already have a noticeable advantage over “naturals” in their life. But the batches of 1000 mean you can now narrow down the specific function of those unknown 60/40 genes. Different batches might have the 60% or the 40% allele, and there would be a patterning scheme. TLDR, you would wait 60 years, and now have the data you need to determine what the next phase of improvements should be. Clone batches eliminate a lot of confounding variables, leaving you with the information you need to pick the “superior” choice for these unknown alleles. (sometimes there wouldn’t be a clearly superior choice, nature has to make tradeoffs, but sometimes there would be)
And do another round of editing and another batch of clones and so on. In 1000 years, you’d have some seriously uber individuals who “naturals” wouldn’t hold a candle to.
Not that I think any of this will happen. AI or robot cyborgs have so much more potential that this is basically a waste of time. Reasonable napkin calculations suggest an AI should be capable of near flawless rational thought at a million times or more the thought speed of humans. They should be so far above us humans in brainpower to almost be the difference between us and insects. Genetically engineered humans would be smarter, but we’re talking “every one of them is as smart as 0.1% genius in today’s world”, which means each one of them is still held back by their biological brain.
This is essentially how the Tay Sach screening program works. Initially it only discouraged carriers from marrying. There is now a bit more detailed testing and counseling, but the net effect on the population wide gene pool is to increase frequency of the recessive allele.
Nearly 30 years ago the cystic fibrosis gene was isolated and characterized. This allowed a line of research which greatly improved the lives of patients with CF with improved treatments. And it also resulted in an increase in fertility of CF patients. Which has the unfortunate side effect of increasing (ever so slightly) the frequency of detrimental CF alleles.
My degree is in genetics. I am quite sure of how this stuff works.
You need to define a metric here. How would you know it works? What result are you looking for? you can certainly change human composition, but how would you know you would improve it?
If you’d given Australopithecus the tools to manipulate genetics do you think Homo Sapiens would have been the result, or do you think they’d have gone for sharper teeth, warmer fur, maybe hands more suited for knuckle-walking rather than tool-using?
Hell, if Ardipithecus had the ability to manipulate genes that whole bipedalism fad might have been nipped in the bud. Walking around on two feet, what will the kids think of next?
Agree with the big picture as presented in your post. But ref the above snip …
As best we understand today, the vast, vast majority of genes do not directly affect the resulting person in any identifiable way. If we were designing a complex system like genetics from scratch we’d have a clear separation of functions where *this *gene controls eye color and *that *gene controls tendency for diabetes or obesity or height or IQ or whatever.
Nature doesn’t work that way. Which is one of the better arguments against so-called “intelligent” design.
Instead almost all phenotypes are the resultant of hundreds of genes pushing and pulling. Everything affects everything else and even teasing out the more significant from the less significant factors is almost never successful. At least at the current state of the art.
Said another way, almost all cases of properties that matter to making uber-humans will fit your second sentence: “In some cases, there won’t be anything you can measure that has statistical validity, if so, just leave it alone.”
Which is why the real eugenics that’s been practiced on domestic animals for centuries is still based almost entirely on watching the animal grow up & breeding the “better” ones together for whatever criteria is “better”. We don’t know, and really don’t care, *how *that maps into the genome. We do know that detail is still far too subtle / obscure for 21st Century science.
It may in fact be so forever simply because of statistics. Absent a perfect mapping of genes and characteristics of many million individuals you simply don’t have a statistically large enough data set to discern the subtle patterns in the diffuse smoke of which hundred factors with which several values add up which way to produce which cloud of results.
Not to mention that genes aren’t destiny. Genes are the first level of control that leads to epigenetics that leads to environmental interaction that leads to the realized individual. Individuals are kinda like tract houses. Yes, there’s a single set of plans for every house of that model in the tract. But each individual house is subtly different because the workers put each nail & board and electrical outlet in place by hand mostly according to the plans but with myriad unaccountable other factors.
All that latter variation in by definition nonheritable. But a lot of the differences we care about are in that nonheritable fraction.
Late add: In the last 200 years we’ve increased the average height of the human race by over 5%. Not by selective breeding but simply by feeding the livestock better. That’s epigenetics and environment at work.
Dogs are not a good example as they are particularly malleable.