Would God's Guidance Be DNA-Detectable?


(S. Joshua Swamidass) #41

There is a controversy between which is best: neutral theory or near neutral theory. For the purposes of this conversation, both do just fine.

Most mutations are not selected, otherwise we’d end up contradicting a massive amount of experimental evidence that most mutations are neutral, and resurrect the problem of Haldane’s paradox. For that reason, even these critics of neutral theory will affirm everything I’ve put here. They would just justify it with a different term than “neutral theory.”

Conceivably yes. Just like this:

However this does not appear to be what we find in nature. So, even though this is conceivable, it does not appear to be the reality in which we find ourselves. Perhaps there is some subtle signal. The obvious signals appear to be ruled out. From neutral theory (or what ever we call it) this shouldn’t be too surprising.


(John Harshman) #42

Why shouldn’t that be surprising? Neutral theory says nothing of what we would or would not expect from divine intervention, only that we expect most fixations to be neutral.


(S. Joshua Swamidass) #43

There is a puzzle in nature. This is one example, but there are many more like it.

  1. Between these species pairs, Humans-Chimps have more functional differences than Mice-Rats.
  2. Between these species pairs, Humans-Chimps have about 5x-10x fewer genome differences than Mice-Rats.

Why is that functional differences do not predict genomic differences? Neutral theory (or variant) gives the answer. Most those differences are neutral, and are just a record of shared history (common descent). It turns out that only a tiny tiny proportion of the observed genetics differences are needed to cause the large functional differences between closely related species.

We can estimate the ratio of selected to non-selected genetic changes. It is something like 1:1000. That ratio, which shows most mutations are neutral, is what makes it clear that:

  1. God need not intervene much to have a large effect on phenotype.
  2. Rare interventions will be lost in a sea of neutral and selected mutations.

Biology need not have been this way. It certainly is not intuitive, and took a lot of effort from Kimura to figure out and make his case. However, given these two findings of biology, we just don’t expect that if God is intervening for the purpose of directing evolution that we would expect to detect a signal for His direction.

If we did detect his involvement, it would be because he somehow wanted to be revealed this way. It would take intentional effort, independent of functional goals, to make it clear in our genomes that He was tinkering.

If biology worked differently, we might end up with a different answer. Biology, however, is what it is.

I do take that back somewhat. Understanding it already, I’m not surprised. However, population genetics is very non-intuitive and surprising. You are clearly an expert here, but that is the fun of this. It is deep enough that it can still surprise the experts. So maybe you do you have reason to be surprised.


Are Humans and Chimps Functionally Similar?
(John Harshman) #44

What is your source for this information?

Well, I can see how “need not” might be translated into “unsurprising”, but I’m uneasy at such a facile assumption of equivalence.

Also, I would not say that rare interventions would be lost; only rare interventions of the same sort as happen naturally at a considerably higher frequency, which is not the same thing. If, for example, God had chosen to cause human chimp differences with a couple thousand microinversions rather than a couple thousand point mutations, that would not be lost in the noise.

Actually, I can’t think of a way to avoid it. Given the size of the human genome, most of it must be junk. And given that mutations affecting genome size happen, I don’t see how that high fraction of junk could be prevented from arising in many species (perhaps dependent on population size).

I don’t say we would expect to detect intervention; I only say that we have no good reason to suppose that we would not. What you have shown is that it would be possible, even reasonable, for God to act in undetectable ways. But you have not shown that it’s the most plausible way for God to act or that it’s the expected way for God to act. Nor have you shown that the purpose for detectable intervention would have to be to make it detectable. You are assuming motives that just can’t be assumed. Your implicit assumption would appear to be that God is parsimonious with means. Perhaps you can back that up, but you haven’t yet.


(S. Joshua Swamidass) #45

A post was split to a new topic: Does Applicability of Math Point to Theism or Not?


(S. Joshua Swamidass) #46

Functional differences is a poorly defined term, but this claim is usually accepted by anti-evolution in regards to Humans and Mice. If you dispute it, go ahead and make your case that humans and chimps are more similar than mice rats. If you want to go that direction, it will surely be entertaining, and we should start another thread.

Perhaps. In that case you have evidentially ruled out that mechanism of guidance. Another strawman down. Keep at it.

I agree with you, especially knowing what we now know. I mean rather that a priori, we did not know if biology would be this way, but it is.

In the same way, a priori you are right. In light of our knowledge of biology, however, not so much.

Yes. That is there, but it is not implicit. I stated it several times in different ways. If that is how you want to put it, go ahead. Honestly, there is some good clarity to it, so I’ll use it too. That is what rules out things like adding history-less chromosomes and such. That would only a parsimonious means if God is trying to communicate to us through DNA. It can’t be justified by the end of creating animals with the “human-condition.”

Thanks for that language. Once again, it is fairly entertaining how easy it is to come to terms with a reasonably resistant atheist scientist. Then I remember the last three decades, and becomes a bit more depressing. @John_Harshman, what we are resolving in a couple hours on the internet, people have dedicated decades of their career to arguing over. Crazy right?


Are Humans and Chimps Functionally Similar?
(Zachary Ardern) #47

Maybe my point about the GP map was not quite clear.

I’m not talking about interventions, but rather the set up of the whole space of physical (biological) possibilities which life walks through. As I see it, there is very plausibly a case for the GP map being constrained in such a way as to facilitate evolution. Whether it is plausible in light of the very best knowledge in the field or not, it is at least conceivable, and I don’t think the neutral theory stuff touches it. If there are large neutral networks which facilitate evolution in our world, but among all possible worlds such networks are rare, then we are facing evidence which would fit nicely in something like a fine-tuning argument, as I’ve written before.

Discovering this would be a form of evidence from DNA. It’s a bit more ‘meta’ than most of the discussion at this thread I think, but still in the realm of empirical DNA-related evidence.

I also think we don’t yet know enough about functionality in the human genome - as far as I can tell, we don’t understand how much transcription is noise and how much is signal, we don’t even know how many protein-coding genes there are in the human genome, let alone what all different kinds of RNA is doing. Estimates from very serious scientists differ four to five-fold (or more) on the percentage of the human genome that is ‘functional’. Defining function is non-trivial and often done badly (I’ve published a phil-bio paper on this - perhaps it was one of the bad ones).


(S. Joshua Swamidass) #48

I disagree. We have fairly tight bounds. Most the disagreement is about definitions.

Case in point. Most of that discord comes from using different definitions of functional.

@Zachary_Ardern, the evidence from human variation is growing exponentially now, and this puts very tight constraints on how much DNA is required for the “human condition”. This information is only growing. We are going to have very good understanding of this just from looking at the distribution of “apparently normal” human genomes.


(Zachary Ardern) #49

Yup, thanks, thought you would likely disagree. The variation within the human population will be informative - but I’m skeptical to what extent. Interactions between variants and widespread redundancy/robustness I think contribute to underestimating function when using this kind of data. Showing that there’s no obvious phenotypic effect with a mutation doesn’t mean that the original sequence was truly functionless - it might be like someone stealing my spare tire… And any definition that makes a spare tire ‘functionless’ is problematic. But we’re not going to really know until we understand the system better (what’s spare/reserve and what’s junk) - with many hundreds of distinct cell types each regulated but doing things differently I suspect we have a lot more work to do.

I think the minimalist selectionist definition of function (championed especially by Dan Graur) is fragile. We don’t have a good complete alternative definition currently, but that doesn’t mean that only 10-15% (max 20ish%) of the genome is functional. There is a lot of equivocation on the word ‘function’


(Eric Michael Holloway) #50

Don’t know enough evolutionary theory/biology to know whether your 1-4 points are valid. But, from a comp sci perspective, it is pretty trivial to figure out what programs are generated by human designers, and which ones are purely a product of a genetic algorithm. Even if the code itself is obfuscated beyond deciphering, the fact we have a piece of software that performs a highly intricate and useful operation is evidence it was programmed. So, it is hard to understand why the same reasoning would not apply to DNA.


(S. Joshua Swamidass) #51

I agree this is usually true, though not always. I can construct a counter example easily. Can you? If you can, you’ve made a lot of progress.

The reason why is because DNA is very much like that counter example. It is just not very much like computer code. We both know the similarities. Great. Start listing out the differences, and it should start to become clear.


(Neil Rickert) #52

Your computer program is abstract; it is disconnected from reality. The DNA is very much connected to reality.


(S. Joshua Swamidass) #53

Good news is that @John_Harshman does know enough evolutionary biology to evaluate those claims. He is an atheist, and is cutting me no slack. Seems like we both agree on the evidential points here. That should build confidence in it, right? Isn’t it great that solid ideas can convince reasonable skeptics like @John_Harshman? Atheist scientists, in general, are very fair.


(Eric Michael Holloway) #54

Not for me. I would like something clearcut. Not relying on credentials, “exercise left for reader”, browbeating, etc.


(John Harshman) #55

I make no case, but I reject your attempt to shift the burden. You are the one making a claim, apparently with great assurance. Your claim may be true. I would just like to see the evidence for it.

You have fairly often seemed to conflate “we don’t know that X is true” with “we know that X is false”. If I were to make a claim about the number of functional differences in humans-chimps and rats-mice, I would claim that we have no good way to evaluate those differences. Perhaps we could if it were a matter of counting functional mutational differences, which is what I originally thought you were referring to. Operationally, it would be difficult and probably beyond our present capability, but I can imagine that there might be a publication or to that would have attempted to investigate such differences. But you have disclaimed that, so it’s even less likely that anyone could do it with a less easily quantified notion of difference.

You consistently claim that any hypothesis contrary to what we actually observe must be a strawman. I reject that idea. When you claim that something must happen, a plausible example that contradicts that claim is no strawman; it’s a counterexample, and you can’t dismiss it.

If you did, it was subtle enough that I didn’t notice it. Can you justify the assumption that God is parsimonious? And if so, wouldn’t it be more parsimonious to dispense with all those billions of years of evolution and just begin where you intended to end up?


(S. Joshua Swamidass) #56

Then you have much studying about evolutionary science ahead of you. It is really worth understanding. Amazing work really. Mathematically beautiful. Empirically validated. Useful for so many people.


(John Harshman) #57

Caveat: so far you have not convinced me (that I can recall) of anything I didn’t already think was true. So I may not be as fair as you suppose. Then again I might be; we just don’t have good evidence yet.


(S. Joshua Swamidass) #58

I suppose it seems self evident. I’m not shifting the burden. Just curious if someone can make a good case against it. None of the creationists I know would. Maybe someone else, even if it is not you, can.

I suppose I do. If we don’t know of if X is true, does not that necessarily mean that we do not know if it is false?

I’d agree. Except to say that from neutral theory we should expect that changes that give rise selectable function should be rare. That is why it is so hard to measure them.

Oh there is quite about there on this. Very interesting stuff! We often post new papers on the forum. Not sure what you think I’ve disclaimed…

I agree that God could have tinkered in a way that was visible to us. I concede that up front. It is also clear that he did not do this. I grant that up front too. You call it a counterexample, but I concede it up front. It is not clear what it is a counter example too.

I like your langauge here and am adopting it. Certainly helps immensely with clarity.

That is not really a problem i my view. I expect here were other purposes in addition to making us that are satisfied by that prior history. This article might make sense to you:


(John Harshman) #59

Some of these claims seem simple enough for anyone. Take the asteroid that caused the K/T extinction. Is there any conceivable way to tell whether it was a random occurrence or was personally directed by the hand of God? Not that I can see, and I don’t think you have to be a scientist in order for that to be obvious. And you don’t have to be a specialist either to see that it had a big influence on the course of evolution. The other three claims may require a little bit more special knowledge, but not all that much. What are your problems with them?


(Eric Michael Holloway) #60

I fail to see why they are undetectable in principle. All I could imagine happening with computer code in a way that is detectable.