Discussing Romans 1:20 scientifically

Peter Rankin

01:13

@JarrodChristman Here are a couple of scientific evidences in nature I see for God's existence to get a discussion started, for which I am unaware of any satisfactory naturalistic reply other than "A solution might exist, we can't rule it out yet," which I don't consider a solution:

(1) The origin of life. I understand the probability of creating a single functional protein randomly, of any decent length, to be practically impossible; often beyond the "universal probability bound," which accounts for exponentially greater than the number of "events" in the known universe, even given billions of years (e.g., less than 1 in 10^150).

The responses I've heard include, "Given enough time, it could happen," but the probabilities are so small that even this is not correct, even presuming the entire universe being a soup of correctly-handed amino acids to begin with, and given billions of years. But the window is far smaller than that, even according to evolution. And what good is a single functional protein anyway, by itself?

The multiverse is another potential explanation, but this is taken on faith, there being no physical evidence at all for it, either direct or indirect. And even given the multiverse, the origin of the first life would be a "freak accident" in our universe, which isn't really a satisfying naturalistic explanation.

I've also heard, "Well, it started with the first self-replicating molecule." But no feasible path is given that avoid error catastrophe in replication, or what kind of molecule this could be; and I've heard of several problems which would inevitably try to tear apart any first self-replicating molecule on a path to life as we know it. It's not a feasible explanation, just a vague speculation.

(2) The design of life. The parallels with intelligent design (say, in manmade systems) is astounding, as positive evidence. And the naturalistic alternative--unguided mutations filtered by natural selection--has a lot of problems that I see.

First, many people think of genetic recombination as evidence for evolution; i.e., the quick results a dog breeder can see when developing a new breed, for instance. But that's almost entirely genetic recombination, or bringing out pre-existing genetic variety, not mutations.

We've never seen mutations generate brand new function of any significance. We've seen mutations break things which can incidentally help--such as the blindness of cave fish, or the inability of a bacteria to digest an antibiotic which is harmful. But smashing things for expedience is hardly a demonstration of mutation's ability to generate brand new function.

And I also think there are lots of problems which make this impossible in theory, even given the large time scales assumed by evolutionists. But that might get into a separate discussion, if you're interested in discussing that.

lupe

09:52

I got suggested this question, don't suppose you're interested in discussing with me? I'd not jump in where I'm not wanted, but if you're interested in a debate I'd be happy to engage in one

I'm a bio researcher, with a strong interest in evolutionary biology and the origins of life, and was sad to see the question locked already when I got here :P

To address the first points, though:
1) The origin of life - this one, is admittedly, tricky. I'm not completely sure of what we'll find here, but I do know that we've, previously, been simply unable to answer the question. And the reason for that is broadly technological. We suspect "the first self replicating molecule forming" is probably a rare event. Rare enough that we probably can't hope to take a vat of chemicals and get something out of it. But, we could, probably, do this on a computer.

Except that we've only had 5 years (I think) of ability to predict some protein structures, and only then through an ML based system, which is not ideal if you're trying to figure out what classes of molecules might be able to self replicate. The thing, there, is, that I question anyone coming up with maths for how unlikely this is.

Because your numbers cannot be accurate, because there's not been a decent search done of possible routes to self replication, because we've not had the tech (and possibly still don't) to do it. So, for now, your guess on if this works is as good as mine, but it still doesn't indicate a creator, just a big glaring question mark.

2) The design of life:
If you'd like an example of mutations producing useful function, well, I'd refer you to the giant pandemic we had, in which we tracked (and I did a tiny, tiny corner of this work, so happy to talk more about this), in real time, the mutations appearing in COVID.

We saw mutations appear that evaded the existing vaccines, made the virus more virulent, and even made it less deadly (which is a great evolutionary adaption around humans - having people less interested in actively wiping it out means the virus gets to replicate). And that's a colossally fast timescale - a couple of years, for major structural changes to the virus. Most things aren't that fast, but add on a few billion years, and you can see how that adds up

Now, you'd probably argue "Oh, but, lupe, this is not a proper creature. Viruses are hardly alive!" - and that's fair. We are, however, going to have a hard time showing mutations on a human timescale in larger organisms, and there's no reason the same maths doesn't work on mice as well as it does on viruses.

But, examples of recent (relatively speaking) beneficial mutations are things like tricromatic vision - we can see a relatively small change between us and apes that gives us the ability to see in three colors. There's a few people with a further modified version, which I happened to find out my partner has while choosing paint colors (but that's a digression :P)

But, finally, my favorite argument: Design shows the mind of the designer. And that's a problem for your position. Because, you see, biology is messy. And, the bits that are not messy are often cruel. The giraffe neck nerve is the classic example - one nerve in giraffes runs all the way down and all the way up its neck. Why? Well, evolution says "giraffes evolved from horse like creatures, and this nerve was fine in a horse, and was never enough trouble in a giraffe

But it's a hard question to answer if you postulate a designer - there's nothing stopping this designer just laying the neck out properly

anyway, sorry, realised I started arguing without waiting for a response

Peter Rankin

11:38

Hi @lupe--sure, that's fine, Jarrod can join in if he would like to as well. To answer your replies:

(1) Origin of life: The problem I see is very strongly theoretical--i.e., error catastrophe, water being an enemy of the first molecule, and functional folding proteins being so "isolated" in protein space, are quite detailed problems with life originating naturally; that these affect basically any postulated path toward life.

It seems like the general responses I've heard are simply extremely optimistically-worded versions of, "We have no idea currently, but we hope to solve it soon." But that's not really a response, since it could be given to practically any argument whatsoever. It's more of a faith position in the powers of science to find a way that, by all mathematical indications, simply doesn't exist.

(2) The design of life: I don't consider these virus mutations to be a good model for generating the design in life. Sometimes viruses "swipe" genetic code from elsewhere. Sometimes their mutations are beneficial because it helps them evade detection, like putting dings and dents in a car might make it look like a different car. It hijacks the machinery of an existing cell to replicate.

But that's not generating new function, e.g., toward new things like wings and muscles and nerves and eyes and ears and brains. If anything, it's simply good at wrecking life. It's fair to say that mutations generating new function in humans would take too long to observe. Yes, and that's part of my point--that evolution via undirected mutations is, at best, unobserved. (And in reality, I see it as far worse than that.)

As for "poor design," I'm not familiar with the details of the giraffe's neck, so I may need to look into that; but in general, here are a couple of responses I would have:

(A) We've been wrong in the past. I.e., we've pulled "junk DNA" back out of the trash can. It reminds me of a five-year-old trying to tell his father how to better frame the house. Maybe in his mind, it makes sense, but he's almost certainly missing something!

(B) Genetic degradation. Since we've been subject to thousands of years of the curse of sin and decay, many things are not quite as originally designed. (Sounds like that doesn't apply to the giraffe example.) But a weathered statue exhibits design, even if it is not in as good a condition as the original.

(C) I don't think God is always designing for maximum functional performance. He also shows His creativity--the way we may build cool structures from only ice. Life wasn't meant to be all about survival in the garden of Eden; there was no death at that point. He may have also wanted to preserve the taxonomic relationship of giraffes with horses (vs. an ancestral connection).

lupe

12:01

I'd sort of concede your problems with the origin of life explanations - there's some good hints at how it came about (some important structures still being composed of RNA, some research showing how fast self replicating structures populate systems, etc)

I guess for me it falls into "God of the Gaps" territory - there's a pretty good chance we'll have an explanation for how this might have happened, and there's not really enough hard evidence either way, leaving it sort of scoreless for science and creationism. I think it's tough, at the moment, to even figure out what the deciding test would be - a extensive search wouldn't prove there's a structure we might have missed that freely self replicates - chemical space is massive

I'd also suspect, conversely, that finding a self replicating structure wouldn't convince you otherwise - after all, it only provides a possible route for life, not the route for life

So we're sort of (absent a major breakthrough of some sort) doomed to this being kind of an unsatisfying answer.

But, jumping into more familiar territory. Going to answer the lettered points first, then switch back to the main argument.
A) Junk DNA is, unfortunately, very much still in the trashcan. There's a minor increase in amount of functional DNA, unfortunately, the often cited "Encode" study took a massive, massive liberty with how it described things as functional (which I'm happy to go into, but it's going to get very bio-nerdy. (but, in short, generating random sequences of DNA and using the ENCODE study's methodology highlights a load of them as functional, which should not happen)

There's no other studies showing the death of junk DNA, and, in fact, it'd be hard to imagine that it would, considering roughly 10% of the human genome is a single self replicating transposon.

B) I'm intrigued by this one - what test would distinguish a weathered statue from a rock? I'd regard it as a tricky argument, as you'd have to show genetic degradation occurs

lupe

12:40

C) I'd argue that, if we can better God's design, and see clear, simple ways to do so, that's kind of a theological problem. Because god is supposed to be all knowing, so we shouldn't really have a hope of being able to improve on this. If we can, I'd argue we have to think again about how it was done - I'm perfectly happy with a theistic evolution explanation, because evolution is this extremely elegant process that makes all these cool creatures. But biology is really, seriously messy

It's beautiful, sure, but so much of it "just works" and by that, I mean in the manner of a car mechanic strapping your exhaust back on with cable ties. All kinds of reactions generate toxic byproducts, rubisco (the key photosynthesis enzyme) is a prime candidate for human redesign, because it inhibits itself in the presence of carbon dioxide, its own substrate. This makes sense in light of evolution, which optimizes for survival, but less when you have an intelligent engineer in the mix.

lupe

13:00

And there's a hundred other examples.

Peter Rankin

13:22

@lupe Interesting, so what percentage of human DNA would you currently consider "junk," and how much of that do you think we may find isn't junk with further research some day?

I looked briefly into the giraffe nerve--I'll have to research that more, but it sounds like the long detour in the anatomy of similar organisms (like humans) can actually be beneficial by providing a sort of sensory input, given the importance of the larynx. If that "detour" makes sense in general, I don't have a problem with God keeping the same anatomy for a giraffe, to show taxonomic relationship, or for similar functional reasons, embryo constraints, or whatever.

To me, it sounds like two people looking at a car engine. The first says, "Wow, look at all this! What amazing design." The second says, "No, look at all these hoses and cords and pipes going every which way, sometimes on crazy detours--no way an intelligent mechanic did this! I could do a much better job myself. Easy."

It seems like a lot of arguments I hear about "bad design" are similar; a sort of naive idea that we could do a better job, while failing to grasp the complexity of it all. I mean, as I understand it, there is so much that we are still learning about biology. Unless we have a solid grasp of all the interdependencies of how it all connects together, aren't examples like the giraffe nerve rather presumptuous on our part?

lupe

My guess is probably the "strongly conserved" estimate - the section of the human genome that doesn't relatively freely change between individuals is about 7-11%
If it changes, relatively freely, it probably doesn't have a function. I am interested to see if this changes with the ability to model DNA in 3d - it's wound round histone proteins, like thread round a load of spindles, so I can possibly attribute a kind of spacer function to some of this. But that would also make insertions of things like transposons absolutely catastrophic, so it's tricky. On balance, I'd say 7-11% in humans. A …

@PeterRankin So, this is less of what I'm saying - it's not that there aren't absolutely fascinating, intricate structures ( the ATP pump springs to mind), but more that there's these odd bungles. It's like someone completed part of the engine, then turned it over to their idiot brother, then went back and repaired all the mistakes.

Peter Rankin

@lupe Just to be sure I understand, are you saying you think 7-11% is junk, or that 7-11% is functional (and thus 89-93% is junk)?

lupe

89-93% is junk, 7-11% is strongly conserved between humans

The rubisco protein, though, is my favorite example - most common protein on earth. Why? because it's the key protein in photosynthesis. It is hands down perhaps the most important protein, too. The problem is, it has a binding site for carbon dioxide that stops it working, so with too much, it halts. The biological patch for this is to produce many, many more copies in a cell than would otherwise be needed, simply to bind as much as possible

Peter Rankin

13:41

So I've heard that DNA has a 3D aspect to it--i.e., it can react to the environment, and "unfold" if you will, to expose the code to generate a protein that is needed to deal with a toxin, as an example. Do biologists have a clear grasp on how much of the genome is used in this kind of dynamic 3D "unfolding"?

lupe

@PeterRankin so, this is a super interesting question (at least for me) - we've got some pretty good hints, though, that it's not going to be a huge percentage. Basically, transposons give us this evidence, they're so called "jumping genes" - basically. integrated fragments of viruses, which can self replicate. These kind of dump themselves into the genome randomly. Now, if they show up in a coding region, everything breaks (as you'd expect).

However, if they replicate into a non coding region, this causes a change in length of that region of the genome. If large sections were involved in precise 3d structure formation, a lot of genes would stop working, from this large change in non coding region length. Yet we don't really see that. Specific non coding regions, that have functions will break. But the big non coding sections seem to be fine with large scale insertions of genetic material

That 11%, by the way, is likely to include things that form a 3D structure - because they're necessary for the function of cells, they'd be less likely to change between individuals

Peter Rankin

16:42

@lupe I'd have to do more research to try to respond to skepticism about the ENCODE project, or about why the rubisco is cited as an example of bad design. So to recap the "poor design" counter-argument, would you still cite the giraffe's nerve as an example? Because to me, it sounds like it's simply biologists not yet realizing the reasons behind the design decisions; i.e., being "in over their head," so to speak, about the giraffe. :)

A lot of "bad design" arguments come across that way to me, actually. Sometimes novices look at the designs of experts and think, "Wow, I sure could do a better job!" Yet the veterans just roll their eyes and realize that it's cockiness. Couldn't that be a lot of it--that we think we know better, when really, we're dealing with systems so complex that we can't begin to wrap our minds around all the potential design considerations involved?

But to a more of an argument on the offensive for design, I'd say mutations filtered by natural selection do not explain the design we see in life. First, they are, at best unobserved. We have never observed mutations creating brand new function, other than getting better at breaking things (sometimes for expedience), which isn't the same thing.

Second, I see a lot of problems with mutations explaining design. One is the very weak nature of natural selection. It does well weeding out serious flaws, but when it comes to small, incremental change, it just seems extremely unrealistic.

I.e., say an organism has a slightly beneficial mutation which gives it a 5% survival advantage, all things remaining equal. But then there's accidental death, variation in the environment, etc. And the massive number of mutations needed to convert one animal to another--assuming taking a direct path from A to B--is enormous. But a highly "detoured" path which add even more.

It just seems like it's not a well-thought-out theory about the origin of design at all. It makes for a good "fairy-tale"-like story, but the details just don't work. And for every beneficial mutation must be many which are very slightly deleterious, and not detected by natural selection.

I heard someone say, "What if we started randomly making changes to a physics textbook. We give half the students the new book with the mistakes, and the other half the original book. Then, depending on which group scores better, we replace all physics textbooks with the new one. What are the odds that this will result in higher-quality physics textbooks over time?"

And that assumes a 100% "kill rate" of natural selection on a group which has a slightly less survival score; but that's not how it would work in reality. Anyway, I just think "mutations + natural selection" is a faith position by evolutionists, because it's unobserved, and I'd also say, contrary to reason.

terdon

17:01

@lupe Can we please stop using the "junk" term? Given the connotations of the word, its use outside the context of the field of genomics and genome study is confusing. We always knew that it was very likely not to be entirely junk, and we now (and for several years) have known that the vast majority of human DNA is actually actively transcribed.

We still don't know the details of how it is used (although I really like John Mattick's theories) but all we really mean by "junk" is "not protein coding".

@lupe There was a really interesting experiment that I saw described in a conference years ago but haven't looked up recently. I remember it was by an Israeli group, and they subjected yeast populations to stress (oxidative, if memory serves) for a few generations and demonstrated both an increase in the overall rate of mutation and an actual whole chromosome duplication -> divergence event.

@PeterRankin DNA does indeed have a 3-D structure and it does need to unfold in order for it to be "read" by the cell. So all DNA needs to unfold before it can be used. @lupe how are you linking that to transposons? Are you thinking of 3D structures beyond the normal tertiary structure? Things like stretches of As giving a kink or long range folding presumably via hbond interaction or something? That sounds interesting!

I find the "unobserved" argument interesting, @PeterRankin. I don't think it's valid at all, since we have actually observed many cases of novel function (both Lupe and I mentioned one) but let's say for the sake of argument that it has never been observed. How is that a problem? The entire basis of theism is about things that have not been observed. At best, the claim is that an old book contains evidence of observation, but nothing can actually be reproduced.

So why is it a problem if you cannot observe mutations producing novel function but it isn't a problem that you cannot observe a god doing it?

Peter Rankin

17:34

@terdon Maybe I missed it, are you referring to the covid virus example from lupe? I don't consider the virus examples to be the kind of mutations required to generate new function, such as creating lungs etc. A lot of times viruses just swipe code from somewhere else, or mutate to avoid detection by the immune system. I'm not familiar with the yeast example you gave, was there a new function cited beyond the duplication of existing code and increased mutation rate in general?

Agreed that merely being "unobserved" wouldn't be enough to throw out the theory. But I saw a sign in a museum once that said something like, "Evolution is a fact, like the law of gravity." A lot of evolutionists say similar things. But I strongly disagree if they mean universal common descent. Even if they mean the kind of mutations which, extrapolated, explain bacteria-to-human evolution, for instance.

So it shows that it's not an "observed science vs. Christianity" for example, the way some (like Dawkins) like to frame it. They're both forensic in nature. So that's my first point--the "unoboserved" just knocks it down a rung from what it's claimed to be.

But then the bigger problem by far, I would say, are the theoretical problems with using it to explain the vast array of design we see in life.

terdon

@PeterRankin I am asking from a more philosophical point of view. I often hear the claim from theists that "we have never observed evolution happening" (this isn't true, but let's say it is for the sake of argument). They then go on to say that this failure to observe it in action is evidence that it isn't real. And yet nobody has seen a creator in action either, so why is this lack of observation not a problem for the theory of a creator but it is a problem for the theory of evolution?

Peter Rankin

@terdon Usually I've heard the "unobserved" comment to prove that both evolution (UCD) and creation are forensic in nature, and so maybe I haven't heard the same theistic arguments you're referring to. My point with that is that it's not science the way gravity is, for instance.

I.e., I think the timescale would too large anyway to observe evolution, and that's been my point in past discussions. Usually I've heard the "unobserved" in response to e.g. Dawkins.

terdon

But, again, why would an inability to observe one thing in action be valid evidence against that thing being real while the same inability to observe another thing in action isn't evidence that the other thing isn't real?

Why is it a problem if we cannot see evolution in action, but not a problem that we cannot see a creator in action?

Peter Rankin

@terdon I wouldn't make that assertion myself. I just use the "unobserved" to place both on the same plane as forensic in nature, and then I try to show that UCD has other far more serious problems.

terdon

Fair enough. In that case, I cannot expect you to defend it.

I will say that you are setting the bar unfairly high though. We can and have observed mutations happening all the time, the only thing we haven't observed live (although we have very good DNA and archaeological evidence for it) is the creation of a novel species (although I suspect we have in the biological sense of "species", we haven't observed the full lineage from say a fish to a terrestrial mammal though).

Of course, this is something that we claim would happen not only over a huge timescale but also something that requires a very large population so it isn't really feasible to test in a laboratory.

But then, the same is true for the theory of gravity: we can only test small subsets of it and cannot, for instance, demonstrate that it can give rise to massive agglomerations of matter that later condense into stars and planets. We can show it on a smaller scale and then extrapolate.

And yet, you state that this theory is somehow "science" where evolution is not. Why do you not require that we also demonstrate that gravity works the way we think it does in another galaxy? Or even another solar system? Why is it only evolution that is held to this extremely high standard?

Peter Rankin

17:56

@terdon If we observed mutations and natural selection working in a way that could be reasonably extrapolated to UCD without issue, then I would grant that it could be considered "science" in the same way that gravity is, at least when it comes to past events. However, I don't believe that's the case.

I.e., the kinds of examples cited (at least with viruses that I'm aware of) aren't the sorts needed to develop lung systems, eyes, etc. I've heard lots of examples of beneficial mutations which are expedient by breaking things (blind cave fish), but that's hardly a step forward, really. We also see rapid speciation per genetic recombination, but that's just bringing out pre-existing genetic variety.

And further, I think the theory of mutation + natural selection is a very naive theory which doesn't take into account all the complexities and weaknesses of natural selection. I.e., it works by killing things (really, non-survival of the non-fittest). And there are only so many organisms it can kill; only so "clearly" it can see, especially w/ micro-mutations. And so it places a massive burden on evolution, contrary to the kind of neat-and-clean idea presented in school for instance.

And, most mutations are very slightly deleterious, and must likely pass "under the radar" of natural selection, and accumulate, while natural selection is supposed to be promoting the positive mutations.

terdon

Sorry, I don't get that last one at all. What's a "micro-mutation"? I do know that proponents of ID make some sort of distinction between something they call "macroevolution" and "microevolution" but that isn't terminology that is used by anyone outside their field so isn't very clear to me.

@PeterRankin Oh no, not at all! Most are simply neutral.

And of course natural selection doesn't promote "positive". There is no such thing as positive. Natural selection acts on things that alter the chances of reproducing. For instance, imagine a mutation that makes you extremely fecund but kills you by age 30. I don't think we would call that positive, but it is something that would likely be selected for.

Peter Rankin

I mean "micro" as opposed to punctuated equilibrium or something, or massive mutations that make very large changes all at once. I.e., how much of a survival advantage would the average mutation in evolution confer? Surely not more than 5% (I'd consider that quite a substantial mutation, actually).

terdon

Oh, I don't have numbers at hand, but yeah 5% would be enormous.

Matthew

@PeterRankin Infinities are weird. You've almost surely heard the claim that, "given enough time", some quantity of monkeys will eventually produce the complete works of Shakespeare. It's not actually true; although the probability given infinite time is theoretically unity, in any finite time, the rate at which they will produce garbage outstrips the rate at which they get closer to producing something useful. In any finite time, the probability gets worse with more time.

terdon

There have been times of massive changes, usually in response to massive changes in the environment. And some really nice experiments that demonstrate that the rate of mutation increases under stress. But that isn't the norm, no.

Matthew

18:01

Infinity is asymptotic; the behavior between infinite time and finite time changes suddenly.

Peter Rankin

@terdon Sorry, I meant a 5% survival advantage. I'd consider that enormous, too (maybe you'd agree?)

terdon

@PeterRankin I would indeed.

I don't want to put any number on this though since I don't know enough about that specific aspect to be able to judge the accuracy of any value.

Peter Rankin

So it seems like that would hamstring natural selection. I think a lot of people think once a beneficial mutation comes along, say a 1% survival advantage, then it somehow gets fixed into the gene pool by natural selection. But that's quite a feat, if it's only a 1% survival advantage, and there's randomness of environment, nature vs. nurture, etc.

terdon

But, 1% is HUGE, why would that be problematic?

We're talking about compound growth here. Even if we assume that the mutation arose in one individual, then that one individual would have a 1% greater chance of reproducing, so would have 1% more offspring, but as we go down the generations, that 1% will very quickly multiply.

Matthew

@lupe Which of those mutations resulted from new genetic information, on the order of more than a few base pair changes? As a counterexample, consider drug resistance in malaria, which took decades to accomplish a coordinated pair of mutations. A coordinated triple mutation seems to be impossible, at least for any reasonable scale.

Peter Rankin

18:04

Well, if you have 10k organisms with a mutation, and 10k without, then after one generation, with a 1% advantage mutation, then even before factoring in other things, that's only a +100 organisms with the beneficial mutation (if my mental math is right, kind of winging it right now). But in a population of 20k, +100 organisms with a mutation for one generation isn't a big deal as I see it.

terdon

@PeterRankin yeah, but this continues increasing down the generations.

Peter Rankin

True, but there's only so much time to get all these mutations in the pipeline and approved before others come along and conflate matters.

Matthew

@lupe There are several reasons the RLN is like that. Development is a big one; adult animals don't simply spring forth into being fully formed, and during growth from a fertilized embryo, the placement is important. "There's nothing stopping this designer just laying the neck out properly" is false.

terdon

@PeterRankin How so?

And what do you mean by "approved"?

Peter Rankin

Sorry, I'm going a little too quickly right now. :) I mean "fixed" into the population.

terdon

18:06

Meh, that just means most individuals carry it, it isn't some stamp of approval or some irreversible thing or anything like that.

Peter Rankin

Yeah, true, I guess what I'm trying to point to is the top speed of natural selection + mutations.

terdon

It's a very slow process, yes.

Which isn't a problem unless you assume a young universe.

Peter Rankin

In school, they seem to present it as if natural selection sees perfectly clearly, or has a 100% "kill rate" for anything that doesn't have the best survival advantage possible.

Well, I think even assuming millions of years, it's still not enough time.

terdon

@PeterRankin Oh, that's nonsense.

It doesn't "see" at all, neither clearly nor unclearly.

Peter Rankin

True, but metaphorically, the teaching is that natural selection is capable of weeding out bad mutations, and fixing beneficial mutations, but I don't think evolutionists consider the massive weight natural selection must carry, even supposing that mutations could by chance "find" the beneficial path to e.g. developing lungs and eyes and brains, etc.

Matthew

18:10

@lupe The problem with the "god of the gaps" argument is that, at some point, you're placing tremendous faith in the ability to bridge gaps that everything we know indicates can't be bridged. At that point, you're not in science territory any more.

terdon

It would only weed them out, as you say, if they made their carrier less likely to reproduce. A classic example is the many mutations we (humans) carry that can predispose us to cancer. Since that usually strikes after the end of a human's reproductive years, natural selection doesn't care one way or the other. So these very "bad" mutations happily remain.

Peter Rankin

I.e., in higher organisms like humans, say the generation is 20 years to be generous--but that adds up really quickly as generations and generations go by, and I don't think it's enough time at all for natural selection to keep the bad at bay (which must occupy a lot of its force) while simultaneously fixing good mutations as well, and with the lack of clarity (say 1% survival advantage, that will take a lot of generations for just that single mutation to fix).

terdon

I don't understand. Why would it have or need some quantifiable force? It isn't a mind, it doesn't think or do. It's a natural result of sexual reproduction.

Ah crap. Gotta run but I'll try and come back later or tomorrow

Peter Rankin

Well, because it operates by death essentially (non-survival of the less-fit), it has only so many organisms it can "kill" to weed out bad mutations, and to fix good ones. And so that's the hypothetical maximum "force," or maybe effectiveness could be another word?

Okay, I have to leave as well...I'll check back though.

Matthew

@terdon The problem with "evolution" is it's such a loaded word. We've observed plenty of natural selection (and "artificial" selection, for that matter.) We have never observed the genesis of significant genetic information. Problems arise when conflating the two.

(BTW, I agree that both sides are dealing with an unobserved past. Creation, however, does have testimony.)

@terdon The standard for UCD isn't "extremely high". The problem with UCD is that it just doesn't fit the evidence well, but it's given a pass because the only other game in town requires a Designer, which is philosophical anathema.

terdon

18:21

@PeterRankin Oh. No, not at all! It does not operate by death, it operates by reproduction. Sure, extremely deleterious mutations can be lethal but that usually means the individual is never born (or hatched or even formed as a fertilized embryo). But the main question is whether or not an individual/group/species reproduces. Death is but one of many ways one might not reproduce.

Matthew

@terdon I'll also agree the whole "micro" and "macro" terminology is... sub-optimal. Say, rather, we see loss or degradation of genetic information all the time, transfer of genetic information sometimes, and new genetic information only very rarely can change two base pairs at once. TTBOMK we've never seen e.g. a totally novel protein, much less novel large-scale features.

terdon

And now I really need to go

Matthew

@terdon How many generations does it take for a 1% advantage to become fixed? How many such changes separate (allegedly) humans from chimpanzees? How many generations are available for all of those changes to become fixed? Those numbers don't actually add up, especially considering the difference is something like 3 million base pairs.

I believe that works out to something like a mutation rate of >100 base pairs per generation. That's not just raw mutations (a value that would be plausible), that's fixation.

Peter Rankin

19:56

@terdon I guess my main point is that natural selection has only so many organisms to work with, whether by "death" literally, or in the sense of a failure to reproduce. I.e., its expenditures on weeding out bad mutations detract from efforts in fixing good ones; and there are so many competing selection pressures, and random environmental variables, etc.

If I get a chance, I might try writing a relatively simple computer program to simulate mutation fixation in a population, with configurable variables for the various "costs" of natural selection, years/days per generation, etc. :) If I can get it done soon and have the time, I'll post a link here, and would be interested to see what you think (and @Matthew). I believe a simulation with realistic settings would be quite dismal for UCD.

Matthew

@PeterRankin Are you familiar with Mendel’s Accountant? (I'm not, especially, but I've heard of it, and it seems relevant.)

Peter Rankin

@Matthew No, I'll have to look that up. Is it supposed to do something similar?

Cool, looks like Sanford/ReMine were part of it, which is great. I may write one anyway that can run in a browser, for the fun of it, that's a little simpler. :) But I may download this one to do a sanity check on my numbers. I'd also like to have a sort of real-time display of mutations, maybe with fun graphs/charts, to make it interesting for non-tech people, and without needing an install. Thanks for that reference @Matthew!

Matthew

@PeterRankin Again, I don't really know. At a cursory glance, I think "yes", but... 🤷

lupe

21:21

@PeterRankin sorry, ended up jumping in a lot later, after a lot of discussion - I think you'd be surprised by said computer program, having written genetic algorithms in the past. (And, yep, an entire field of research/programming, that uses evolutionary logic, and is extraordinarily effective

en.m.wikipedia.org/wiki/Evolvable_hardware - Adrian Thompson did the classic work on it in the 90s, basically setting up a genetic algorithm to program a big load of programmable logic gates

What's really, really cool about this example is it illustrates pretty much the same "screwy design" we see in evolution - the final result of his circuit functioned, but when you removed a few non electrically connected cells, it stopped working - so it was using some kind of electrical interference with the non connected cells. Thompson, an electronics engineer, was quoted as saying "well, really, I don't know how it works"

@Matthew problem with this maths is mutations happen in parallel. So each gene that has a mutation is under selection pressure each generation

Matthew

@lupe Even so, does ~100 mutations becoming fixed per generation seem reasonable? (I don't actually know, that's why I'm asking.)

The Guy with The Hat

just to give a rough mental model of exponential improvement, if some form of life had 1 generation per year and got 0.00000001 better each generation, after 4 billion years it would be 100000000000000000 times better

those numbers aren't based on anything in particular, I'm just illustrating how much of an effect compounding could have

lupe

@Matthew 100 mutations over (3 billion base pairs* number of organisms in the population) seems fairly reasonable

Possibly a bit low

That's for humans, and we're relatively small, as genomes go

Matthew

@lupe Not just occurring... becoming fixed in the population.

Eph

Interesting the Mendel's Accountant program seems to assume that mutations are either beneficial or detrimental with nothing in between. I'm pretty sure that the vast majority of mutations do nothing, so that doesn't seem to fit reality very well.

Matthew

21:36

Anyway, a general problem with all this is that organisms are, for the most part, getting less healthy over time.

lupe

@Matthew fixation in genetics is where that becomes the only allele in the population? Why is that part of this criteria? Most populations exist with a pretty massive mix of alleles

Matthew

@Eph I don't know if I'd agree with "do nothing". A significant number are, at most, minimally harmful.

lupe

@Matthew this ignores the aggressive selection against this

@Matthew there are plenty of harmless mutations - swap some base pairs, and the amino acid coded for will be the same. After a gene duplication, you have two copies of the gene. Mutations are pretty free to act on one of the copies - after all, there's a functioning copy, so at a certain point, those mutations are functional but totally harmless

Matthew

@lupe Let me see if I can clarify by example. To start, I have one population. After 10,000 generations, I have two populations whose "average" genome differs by 2,000,000 base pairs. Somehow, 100 mutations per generation have to "get together" such that all 100/generation are spread throughout the "final" populations. Is that 100/generation number reasonable?

@lupe That isn't necessarily true. Some codons that specify the same the same AA affect folding.

lupe

@Matthew how?

@Matthew that maths seems, honestly, completely reasonable.

Matthew

21:43

@lupe I haven't looked into how...

lupe

For a start, only 1 in 10 of those mutations will be in a region that does things

So we're down to 10 mutations per generation, assuming they're spread randomly. However, they're unlikely to be - many gene mutations are lethal, selecting them out of the population, so call it, back of the envelope, 2-5 in 100 that fall in a region of the genome that codes for a protein.

Matthew

Hold on, I do remember... something about the relative abundance of the relevant... tRNA? (The thing that binds to the codon in order to attach the amino acid.) Different codons change the timing, which can impact folding.

lupe

Sorry, close to a region that codes for a protein, the conserved 11% of the genome is not all coding

@Matthew ah! It's unlikely to impact folding! (Sorry, this is why I had a blunt answer, it didn't make sense). What it will do is influence the relative abundance of the protein.

Ooh, wild, ok, this is cool, my apologies - I can see how that works now

Matthew

@lupe That's not how I remember it explained, but even if so, changing the relative abundance of the protein can be problematic too.

lupe

@Matthew sorry, I looked it up quickly, you're right, for some proteins

Matthew

21:50

Hey, we can both be right. 🙂 Controlling the relative abundance of a protein is also a meaningful task.

lupe

It's not a super common effect, but it exists. So, changing relative abundance of proteins would be bad, except that cells have feedback loops for that, that controls protein abundance

Eph

interesting! In the same paper published by the Authors of the Mendel's Accountant they say "The reason these lines remain straight and do not change
their slope significantly when selection is added, is
because most mutations are nearly-neutral in effect,
and are consequently immune to selection. Only the
worst (or best) mutations can be selected, and all the
rest continue to accumulate in a linear fashion." So they're using the fact that most mutations are neutral to justify not selecting on them strongly, but then calling them detrimental.

lupe

What there is is selection pressure towards common tRNAs, though

@Eph yep! This is precisely the problem with the maths. So, it takes a fairly massive selective advantage to rise above the noise in genetics

Peter Rankin

@lupe But isn't that the issue? Most mutations which are only slightly detrimental are not "weeded out"; and if most non-neutral mutations are slightly detrimental, then wouldn't that accumulate in the genome quietly, like rust accumulates on a car?

I.e., I heard an illustration that if we introduced type-o's into a physics textbook, and then had half the students use the original, and half use the one with the mistakes, and scored them at the end of the year, and then chose the book whose students scored highest--would that really end up producing better physics books over time? Most of the changes will be practically negligible taken alone, but over time, it will destroy the book.

Eph

That's not a very good example though. English typos are very different from genetic mutations.

lupe

21:59

@PeterRankin I have a good response here, but have to feed the cat first

Eph

Most mutations literally do nothing different. Most typo's break word making it harder for a human to interpret but still parsable. But there's no human to think and interpret a typo differently in DNA so that it's only slightly worse. If it's going to make something worse, it'll likely prevent a protein from forming "correctly" so that the protein doesn't work like it's "suppsoed" to. That will likely get selected against.

Peter Rankin

@Eph Do you have a link to what you would consider a correct plotting of random mutations (for say, a given organism), which plots their harmfulness/benefit (or neutrality), plotted by probabilistic frequency?

Eph

Sadly no, I'm very interested in one though. I was excited to try out the Mendel's Accountant until I realized it's contradiction of non-selectable yet detrimental.

Peter Rankin

22:15

@Eph Well, do you think it's reasonable to say that a great many mutations have less than, say, a 0.1% selection detriment? If so, that seems like a pretty good chance of it becoming fixed in the gene pool eventually, with all the other mutations that may be thrown into the mix? Is there really much difference between that and, say, no selection pressure on that mutation at all?

And if, suppose, 0.1%-selection-detriment mutations are common enough, wouldn't they be able to "slip in (nearly) unnoticed" by natural selection? So yeah, the really bad mutations would be eliminated, but the little ones could still accumulate quietly, like "genetic rust"?

Eph

I'm not familiar enough with the numbers to say for sure, but no. If it doesn't significantly get selected against, then kind of by definition it's not really detrimental.

Peter Rankin

But isn't the degree of selection pressure conditional on the amount of detriment a mutation has? And so wouldn't a slightly detrimental mutation have only slight selection pressure against it?

Eph

So, sure there's going to be accumulation of non-detrimental mutations, that's the whole theory of evolution. But they aren't detrimental, so calling them rust doesn't really make sense.

Peter Rankin

So leaving aside truly 100%-neutral mutations, do you have a rough idea of how many mutations will be beneficial vs. detrimental, by percentage? It seems very unrealistic to have a 50/50 split in my view, given that almost all changes to a highly specified system are almost always negative in some way, big or small.

Eph

22:30

I think that's a great question! I'd guess that the majority of the non-neutral mutations are probably negative, but that they will be selected against in some monotonic relationship to their benefit. In fact, that's probably how I'd measure how good or bad they were.

If they are selected against, then they're negative mutations. If they're selected for, then they're positive mutations. If they don't have any selection pressure, then they are neutral mutations.

Peter Rankin

So do you think that slightly-negative mutations are unlikely (or even impossible)? I.e., I know it's different than English text in ways, but it still seems like slightly-negative mutations are very possible? Or even the norm? That's part of the case Sanford makes. I'll have to check his book to see the support he gives for that again. But it certainly seems intuitive.

Eph

I think it seems intuitive if you imagine that the evolution is trying to achieve some goal besides reproduction. Like with the book example, it's trying to convey some information and get some kids to do well. But with evolution there's no goal besides reproducing. So sure there's going to be detrimental mutations that reduce the lieklyhood of reproducing, but, by definition they will weed themselves out. Anything that's detrimental enough to cause extinction will, and that species will be gone

Peter Rankin

I'd agree that selection pressure would be proportionate (roughly) to detriment or benefit. Maybe what I'm really asking is, how frequent (if at all) do you think that slightly detrimental mutations occur? Because if a sizeable percentage of mutations are too slight to be effectively eliminated by natural selection, then certainly the species could deteriorate very quietly, like rusting out. That's Sanford's argument, basically.

I.e., lots of mutations with a detriment of 0.1% (say), if the volume is high enough, could certainly overwhelm natural selection's ability to weed them out, and thus they'd slowly bring the species down, wouldn't they?

Eph

22:49

I think that's probably true. If you have a high enough mutation rate, and the distribution of mutations is strongly skewed enough towards failure to reproduce. Then indeed the population would die out.

lupe

I think the problem with this line of reasoning is that it assumes a 0.1% detrimental mutation combined with a 0.1% detrimental mutation equals a detriment of 0.2%

I'd argue that's not often the case - imagine a mutation that, say, makes your muscles need 1% more energy to work. Then imagine a mutation that means your cells produce 1% less energy - what's the practical result of combining these two?

Peter Rankin

Well, isn't evolution assuming that slight benefit plus slight benefit equals new life forms over a long time? It seems like that must also imply that slight detriment plus slight detriment, if the case, must equally lead to decay.

lupe

@PeterRankin sure - but combinations matter - imagine a rabbit that gets 6 beneficial mutations

Eph

That's a good point, loosing one copy of a gene when you have 2 might be detrimental, but loosing the second copy is a non-starter, so you'll never see someone missing both copies. And the more people who loose one copy will create stronger reproductive selection for people with 2 copies.

lupe

The rabbit ends up breeding lots , spreads all 6 of those at once - it has outsized benefits to having combinations of beneficial mutations

Peter Rankin

22:57

But surely a consistent flow of slightly detrimental mutations, if overwhelming natural selection, must lead to decay, right? Another ding on a car might not reduce its value noticeably (if at all), but enough of them surely will destroy it eventually.

lupe

@PeterRankin but you're not just dealing with one car, though. You're dealing with a population (and, note, this only holds for decent sizes of populations, the effect you describe absolutely happens in, say, cheetahs, which are currently undergoing collapse)

Sorry, cat walked on keyboard

Peter Rankin

Well, maybe that all comes back to getting what you consider a realistic graph plotted by benefit and detriment, with probabilistic frequency. Sanford gives one that makes a lot of sense to me, and in his model, most are slightly detrimental.

@lupe An example of a clearly detrimental mutation in English, courtesy of your cat. :)

lupe

@PeterRankin an interesting study here pnas.org/doi/10.1073/pnas.1313424110 on delitirious mutations serving as stepping stones to advantageous ones, which makes sense if they're not very delitirious

But, yes, it's an interesting piece of dynamics - I might ask my old boss if he's got a good paper on this (statistical evolution is more his thing than mine), and might try some agent based modelling, here. My suspicion is that delitirious mutations accumulate when you model it as a simple system, but anything more complex causes them to fall away

Partly, because we don't see constant breakdown of viral or bacterial populations - we'd expect them to experience complete collapse very rapidly

We definitely didn't see that in COVID, which I'm pretty familiar. So, this means the Sanford model has an error - it doesn't match up with observations

I'm not completely sure why, but I think it should be possible to figure out. This is why I like these discussions, by the way! It's good for my knowledge, too!

But, yes, if this was true, the species with the fastest growth rate would just collapse completely.

Transcript for

Discussing Romans 1:20 scientifically