Evolution, subtopic = mutation (GQ and GD content)

Great Debates
1

In this thread I would like some help teasing out the thinking & visceral-dislike reactions that cause some people to reject the theory of evolution. I’d like to bracket off and ignore the ones whose reasoning amounts to “My pastor said God created the world in seven days” or “I didn’t descend from no freakin’ ape-man”, and if, in your opinion, that doesn’t leave much, feel free to say so.

I’ll toss in some of my own mental processing on the subject matter, along with the reevaluation stuff that led me to get over those objections and embrace evolution theory as an accurate model of life differentiation, and then follow it up with some GQ questions on some of the relevant particulars. I am thinking probably my own mental processing might overlap that of other folks who express misgivings about evolution, and if that is so, perhaps that suggests areas of the theory that need to be explained to the general public with more clarity. From the thread-title, y’all probably already anticipate that the bone of contention lies with mutation rather than the natural selection thereof —

OK, as taught to me (haphazardly) in science classes in grade school & jr high, you have organisms of an existing species and during reproduction the DNA code that tells the new organism how to develop — the blueprint —is duplicated but sometimes random errors occur during the transcription process. Then natural selection weeds out the variations caused by these random errors if those variations are non-advatageous, leaving behind the “normal” individuals who exhibit no such variations but also leaving behind those individuals with advantageous variations, whose DNA blueprints may be passed on to be expressed by subsequent individuals. If the variations are sufficiently advantageous, the previously “normal” individuals will lose out under natural selection and the population will all come to exhibit the variation after awhile, and that’s evolution.

The natural selection part of it is so self-explanatory it’s hard to wrap my mind around anyone rejecting it. But the mutation part…

Let’s say I open up my copy of Microsoft Excel with a hex editor. I randomly choose some blocks and randomly pound down some keys to overwrite some passages with random errors, then save the changes. Launch Excel. I figure likely outcomes are:

• It launches normally and when I try to use it it acts as it always did. Whatever I overwrote with random keystrokes is in some part of its code so seldom called upon that in normal use I may never see any difference;

• It launches normally and behaves normally until I try to perform Function X (let’s say: select “Save As” from the File menu). When I do that, Excel crashes, repeatably and dependably. My random changes have given Excel a disability although it still runs.

• It launches normally and behaves normally until I try to select “Save As” from the File menu. When I do that, the Save As dialog box apepars drawn all out-of-kilter. Icons, dropdown menu widgets, the text-entry field where you input the file name, are all either missing or don’t react when I click on them or make incomprehensible screen artifacts appear when I click on them. With a little experimentation I find that I can actually perform a “Save As” with this interface, although I have to do it in a different manner than from a stock, unaltered “Save As” dialog box. My random changes have made Excel different in a certain area, and it could still be regarded as a disability even though it still retains functional purpose.

• Excel won’t launch. Whether because code necessary for Excel to run has been destroyed or because an internal self-checking mechanism (checksum for instance) stops the program from executing because of contradictions and inconsistencies, my modifications have killed Excel. Fatal variation.

With me so far? Well, what I always figured I would not see happening as a consequence of my random-string keypounding would be anything where suddenly the “Save As” dialog box has new features, improvements, etc. With extremely rare, statistically off-the-page kind of rare, you-could-do-this-until-the-stars-burn-out-without-it-occurring kind of rare exceptions, all the changes I cause to occur in Excel are going to be described in terms of how much damage I did. So here’s the theory of evolution telling me that we start with some carbon-chain molecules and a zap and get a combination that self-replicates (believable), that random variations (believable) occasionally yield advantageous differences that are selected for over time (ummm…) eventually causing sequences of changes producing salamanders, pine trees, tyrannosaurs, blue lichen, blue whales, and wildebeests (your’e shitting me?), which are subject to natural selection weeding out the less well-adapted (very believable).

GD Question: Is this general train of thought well-represented among folks who reject evolution as a plausible model, and if so, would elaborating on and explaining the mutation part of the theory more explicitly help address some of that?

My own reevaluation stuff — For me, it was seeing a NOVA 60-minute special about “Chaos”, followed in short order by me reading the James Gleick book on the same subject. Aha, I said, there is order emergent in turbulence, in apparently random behavior! Rather than the DNA replication errors being absolutely patternless random noise in the signal, they could be examples of a dynamical system that normally does x, y and z (or a, t, g, c, if you prefer) but at unpredictable intervals which nevertheless may exhibit a pattern does a variant, and the variants produces also follow some kind of pattern, aggregate in some way around some kind of “strange attractor”, exhibit some kind of underlying order even as they remain unpredictable! All right, I’m not a physicist, a biochemist, evolutionary theorist, etc, and this may have no direct bearing, but it knocks me off my “random noise can’t be generating all this complex order” platform.

GQ question: Would anyone who feels they have a decently solid understanding of the current scientific thought on mutation care to comment and elaborate on any of the above w/regards to how mutation can product useful variation rather than useless noise enough of the time to account for speciation? Thanks.

2

First of all, never think about DNA as a “blueprint.” A blueprint implies that every element is laid out and specified in some recoverable form. But it’s not. DNA builds one way: it actually describes a process, and while you get a fairly predictable outcome from carrying out the process, we’re talking something much more like a recipe than a blueprint. You can’t unbake a cake like you can dismantle a car.

This has some pretty big implications for how mutation affects outcomes. What you are working with is not the code for Excel, but rather something that describes how to go about producing code that will function roughly like Excel. If you duplicate all the code in Excel and compile it, you’ll get lots of function errors from crossed variables, doubly-defined functions, and so forth. But if you duplicate the code for producing Excel’s code, you might just as easily get two Excels.

But even these are actually some pretty terrible analogies. The problem with biology is that there are no good analogies. Nothing else is anything like it, so you can’t take shortcuts to understanding how and why it works. It’s incredibly complex. For instance, proteins don’t have to do just one thing. Because the logic of proteins folding only operates with regards to certain very specific sequences, proteins can have a lot of arbitrary nonsense in them that doesn’t affect their shape (and hence, doesn’t affect their function) much at all. And yet these redundant parts can then develop functional traits of their own that don’t affect the folds of the relevant areas of the previous function, leaving the protein now capable of doing two entirely different things.

Likewise, you can duplicate proteins in DNA without having the end result be double the amount of proteins (since although there are now two different sections making the same protein, the same regulators operate to keep both in check and the population size the same). Then, one of the proteins can start to change a little, making it less subject to the regulation and having a different function: the result being that the old functional protein is retained, while a new one emerges, allowing experiments with new functions that don’t break the old ones.

All of this is pretty unique to biology (though we can model it with computers now), and there are so many other quirks like this to learn that you can’t really judge whether the system is plausible or not until you learn most of the ones we know about so far.

Scientists often try to talk to people about this by simplifying it and using analogies, and that’s okay if that’s all you’ll ever need. But for a layperson trying to get a deeper understanding of biology, these very simple and often incorrect images often make them think that they’ve discovered problems with the theory that are actually just a result of their very limited understanding of what’s really going on. I think that’s a big part why anyone with an inclination to want to reject evolution, and who tries to get a deeper understanding of what they hope will be weaknesses, find so much to support their skepticism. They fall very easily into the trap of having a little bit of knowledge, but not enough to either let them know that they are lacking, or that they are drawing wrong conclusions. For instance, as a recent post on Pharyngula notes, some creationists think that to evolve an eye, you’d have to evolve just one eye, and then somehow duplicate it, so evolution should show a lot of one-eyed creatures (and it doesn’t, so its false, yadda yadda yadda). But the opposite is true. Because most eyes evolved after bilateral symetry was laid down, it’s actually much much harder/rarer to evolve non-bilaterally symetrical patterns (like having an eye on only one side and then duplicating it on the other side). Slipups and counter-intuitive elements like this abound.

What’s the answer to this problem? I dunno. Biology just is hard. At some point, we can’t really simplify AND be accurate. But for educating most people, the simple take is the only thing people have the time and interest for.

3

First of all, as far back as 1959 a researcher at IBM experimented with constructing programs through mutations - what would be called Genetic Algorithms today. So, software through mutation is not impossible.

Second, you underestimate the number of experiments performed. I think there is an average of six mutations per person, and with the billions on earth, even humans conduct a lot of experiments - not to mention bacteria, etc. Say, in your example, a bit flip changed the default number of files to show in the history part of the File Menu to something bigger. Might be advantageous - and possible. With 30 billion experiments, who knows what you’d find?

Another difference is that smaller changes in DNA cause bigger changes in the individual - like fruit flies that grow new legs with a “single bit” flip. One of the bugs we worry about in testing memories is single bit addressing faults, which cause the computer to read an entirely different column of memory from what was intended.

Could a big problem be that some think we’re perfect, and so don’t accept the possibility that a mutation might improve us?

4

Yep, that’s about it.

I’d recommend the following book, which I’m just in the middle of myself. It addresses the specifc questions that the Op is asking. You’re not going to get someone to explain this to you in a short post on a message board.

Check out The Plausibility of Life; Resolving Darwin’s Dilemma by Kirschner and Gerhart. It’s not the easiest read in the world, but I don’t think you’ll have any problems following it.

5

Ug. Even biologists seem to get this wrong. Phlyetic gradualism, or continuous steady change, was never a dominant view for very long, and hasn’t been for a long long time. Furthermore, even Darwin didn’t propose it. In Origin, he several times says that he doesn’t actually imagine that real world change is steady or continuous, he’s just using simpler diagrams and models to explain the basic concepts.

6

Yes I’ve seen that argument put forward by IDers… And the reason the Excel analogy falls down is that the code is static and unchanging (other than the mutation). This is why hills don’t evolve, they “mutate” but they do replicate themselves, so the don’t evolve. Living organisms are constantly recombing their dna with other organisms, producing new organisms. So even without mutation there is a great deal of variation in genetic material (and unlike mutation it is not completely random, being dependent on choice of mates). Once you have recombination AND mutation you have the essential mechanism of evolution.

7

The problem with your computer analogy is that a computer program is specifically designed to be as advantageous as possible. An organism isn’t designed. So a computer program would be impossible to improve through random mutations because if it were beneficial, it would have already been added by the programmer (the exception being a benefit that the programmer never thought of).

8

Well, the environment of a computer program does change. Say Excel originated on an old PC with scant memory and disk. The max number of rows would be relatively small. If it was moved to a more modern PC, a mutation increasing the number of rows would be advantageous. If this mutation had happened in the original environment, and say resulted in tons of memory being allocated, it would be disadvantageous, causing the program to crash or run very slowly.

So, even if there was a designer 100 million years ago, the fitness of creatures in today’s world would still need to be explained.

9

More rows would be advantageous to the user of the Excel program but why would it be a disadvantage to the program to have a severely limited number of rows?

10

In that analogy, the number of users the program had would be like its reproductive success. Fewer rows= fewer users (fewer offspring).

11

So there would be a line in the program that erased it if the number of users was too low? In other words what would be the harm to the program if it had fewer users?

12

None, if you look at it that way. But then it’s not a good analogy. To make the analogy useful, there has to be something analogous to reproduction. Usually you postulate some sort of regeneration based on the success of the original. If you assumed the reproductive success was analogous to the number of purchacers for a program, then that would be the “harm”-- not many people would buy the version with fewer rows.

If you want to assume there is only one version of the program available, then it’s again a useless analogy.

13

It would have to reproduce in some way, which I’d guess would be people copying it for their PCs. If you got several versions, you’d take the one with the most rows that fit, and would never send the one that crashed to your friends - in fact you’d probably delete it.

And yes, this analogy is getting rather strained.

14

It was strained to begin with. As I noted in my post, any real discussion of biological evolution cannot use analogy. The point of analogy is to simplify concepts using common experiences. But the complexity of biological evolution isn’t anything like anything else in anyone’s experience. The only things like it in existence are computer models of it, and even these are so complicated themselves that we have exactly the same problem: to understand them, you have to first spend tons and tons of time figuring out their unique complexities. You can’t get there by analogy alone.

15

It’s interesting that the basic concept of evolution is so simple - descent with variation, filtered by reproductive success. But the details are in trillions of experiments in hundreds of millions of years.

As an aside, those who say we can stop doing animal experiments and use computer models instead clearly have never actually written a computer model.

16

Indeed. We are actually only now, on the best super-computers, just starting to be able to model protein interactions to get a sense of how viruses work. And even those models involve a fair amount of fudges.

17

Yeah, it’s been around forever in one form or another, often involving even worse analogies like firing bullets into TV sets.

Not as far as I can see. The problem is that we are discussing an aspect of science that some people have a pre-existing desire to reject. They aren’t looking to understand or evaluate or even to critique, they simply want to find flaws. If you elaborate on one bit to the point where it becomes so comprehensible as to be undeniable they will simply move to the next point. It’s just a manifestation of the God-of-gaps aspect of intelligent design. And as others have pointed out this is a field people spend their entire lives studying and still find flaws.

As others have pointed out, it’s sufficiently complex that you are never going to be able to teach even the main points to a level where misunderstanding is impossible without several years of intensive study. Or to put it another way only people with a deep interest and a genuine desire to lean will ever take the effort to evaluate the major flaws, and those people make up such a trifling proportion of those who reject evolution that addressing their needs is like pissing in the ocean.

As others have said, you chose a poor analogy. If we restrict ourselves to the most common mutations then we are talking about altering or at worst removing single amino acids. Those amino aids are only small parts of the entire protein and are unlikely to render the whole thing completely useless. We can see this in conditions like sickle cell where single point changes result in a working copy that has different properties but by no means useless.

For your analogy to be even vaguely accurate you would have to replace or remove entire words, not simply random letters in the middle of words. And those words would have to have the potential to make some sort of sense in context. You couldn’t simply replace “push” with “bulldozer” That better reflects the effect of mutation, though it is still far from perfect. So rather than replacing “switch 10000” with “jskkwkkitykch 10f0g00s0s” you would be obliged to replace it with “push 1001” or something to that effects.

If you even make those minor changes to the analogy you will start seeing that things aren’t as impossible as they appear. Very few changes will result in a totally unworkable application. Usually you will simply end up with dead functions. Often a keystroke combination that should be “print” will become “save backup” or something like that. Just occasionally you will find the new combinations better for what you do and stick with them.

But really it was a bad analogy to start with and can’t be easily salvaged. DNA and life generally isn’t an unforgiving binary code. It is more analogous to handwriting where even fairly major variation between individuals can occur without obscuring the meaning or functionality.

18

There’s more problems than just this with the analogy. It is based on a fundamental misconception about how evolution works.

It is better to think of it this way. One has a simple program on the computer, with a few functions – store numbers, add them, multiply them, etc. One of the functions is to add two numbers, and overwrite the first number in the first register with the sum. Instead of one copy of the program, you have 20 million copies of it. The program has a lot of extraneous code in it, and each of the 20 million copies have differences in this extraneous code as well as the code that actually controls the key functions. This extraneous data can sometimes be copied from other programs, sometimes from the same program, sometimes just random junk. But here’s a key: all of the programs funciton perfectly with regards to their core functions. Let’s think of these as the functions necessary to “reproduce” and be propogated.

Now, let’s add to the selection criteria a bit. Something small to begin with, like instead of adding two numbers and placing it in the first register, it adds two numbers and places it in the third register. Only 5 of the programs of the 20 million do this. So we take the 5, we copy them out to 20 million copies again, and make the extraneous code varied again. Now we have 20 million copies, all slightly different, that add two numbers and place it in a third register, as well as the core functions.

Lather, rinse, repeat. The selection criteria are selected semi-randomly for a few billion generations. There is no goal here. Many criteria are retained, sometimes old criteria are dropped. Now after let’s say ten billion generations, we take a selection of the most complicated programs. By this point, they are huge, horribly programed, behemoths. But they cope well with a huge variety of different criteria.

Note that this is not a perfect analogy either. Most of the other contributions to this thread are equally correct. But introducing random mutations into Microsoft Excel is not gonna cut it in terms of analogies.