Why Evolution is a Fairytale for Grown-Ups

[scruffy]

I have several fields I can discuss fluently. What you talk about is not my field.

Ah - but it is, you just don't realize it yet.

We're talking about evolution. And I'm saying that any time you get a generator + a recursive process, you get evolution. All self replicating processes evolve, by definition.

The peculiar aspect of "biological" evolution is its variability, which in the self replication world equates with "mistakes". The biologists call it "mutation".

Mathematicians study two types of mutation: changes in the data, and changes in the process. In physics, changes in the process are rare, because you have physical laws. In biology this constraint doesn't much apply, because the process is independent of the underlying physical laws.

However in each case, changes can be described in terms of random variables. This is where information theory meets biophysics. You can formulate these issues like brain teasers in statistics.

For example - let's say I have three bags, and three different color marbles (let's call them red, white, and blue for convenience). I put some number of each color marble in each bag. If I then choose a bag at random and pick a marble from it, what are my chances of drawing a white marble?

To answer that, you have to be informed of how many of each color marble are in each bag.

So, if I have 4 different nucleotide types and I put them together in groups of 3, and each group codes for one of 24 amino acids, how many different ways can I fold a protein of length 20?

To answer that, you have to be informed of all the possible bonding angles between neighboring amino acids - because you have to know which ones are equivalent. Maybe the sequence ala-thre-pro is exactly equivalent to the sequence ser-thre-pro, and therefore a mutation in the first three nucleotides is meaningless and has no effect.

Now consider this variation: every time I draw a red marble, I put an additional white marble into one of the bags at random. That's a recursion, and it will cause your set of partitioned marbles to evolve.

In DNA biology there are insertions, deletions, and substitutions. So you either put a marble in, take a marble out, or take it out, change its color, and put it back in. You also have duplications, where you take an entire bag, duplicate its contents into a new bag, and add the new bag to the pile.

So data (marbles and bags), and process (operations). The goal being for the end result (protein) to have a certain shape. In computer science they call these "evolutionary algorithms".

 

[Robert W]

Ah - but it is, you just don't realize it yet.

We're talking about evolution. And I'm saying that any time you get a generator + a recursive process, you get evolution. All self replicating processes evolve, by definition.

The peculiar aspect of "biological" evolution is its variability, which in the self replication world equates with "mistakes". The biologists call it "mutation".

Mathematicians study two types of mutation: changes in the data, and changes in the process. In physics, changes in the process are rare, because you have physical laws. In biology this constraint doesn't much apply, because the process is independent of the underlying physical laws.

However in each case, changes can be described in terms of random variables. This is where information theory meets biophysics. You can formulate these issues like brain teasers in statistics.

For example - let's say I have three bags, and three different color marbles (let's call them red, white, and blue for convenience). I put some number of each color marble in each bag. If I then choose a bag at random and pick a marble from it, what are my chances of drawing a white marble?

To answer that, you have to be informed of how many of each color marble are in each bag.

So, if I have 4 different nucleotide types and I put them together in groups of 3, and each group codes for one of 24 amino acids, how many different ways can I fold a protein of length 20?

To answer that, you have to be informed of all the possible bonding angles between neighboring amino acids - because you have to know which ones are equivalent. Maybe the sequence ala-thre-pro is exactly equivalent to the sequence ser-thre-pro, and therefore a mutation in the first three nucleotides is meaningless and has no effect.

Now consider this variation: every time I draw a red marble, I put an additional white marble into one of the bags at random. That's a recursion, and it will cause your set of partitioned marbles to evolve.

In DNA biology there are insertions, deletions, and substitutions. So you either put a marble in, take a marble out, or take it out, change its color, and put it back in. You also have duplications, where you take an entire bag, duplicate its contents into a new bag, and add the new bag to the pile.

So data (marbles and bags), and process (operations). The goal being for the end result (protein) to have a certain shape. In computer science they call these "evolutionary algorithms".

I own a university book about evolution. I have it next to me. I shall investigate the book and see what it says about colored marbles.

 

[scruffy]

I own a university book about evolution. I have it next to me. I shall investigate the book and see what it says about colored marbles.

I have a university textbook on phrenology.

So what?

Do you believe everything you read on Google, or can you think for yourself?

 

[Unkotare]

You get to walk slowly through this scientists discussions where his aim is to prove evolution is an adult fairy tale.

Be careful since he just may persuade you that evolution as taught is an adult fairy tale.

Pretty sure this should go in the religion forum.

 

[scruffy]

The exercise with marbles is called Bayesian statistics.

The issue is, how much information did you have before the change, and how much information do you have after the change?

(Question for extra credit: is it possible to lose information from a change?)

If there is a substitution, some of the probabilities change.

If there is a duplication mutation, all the probabilities change all at once.

There are machine learning algorithms that handle these cases. They work on maximizing and minimizing probabilities.

Much of the protein folding AI used in cancer research and drug design works this way.

It is impossible to discuss evolution in scientific terms, without being aware of these mechanisms.

Biologists are generally not very good with math (although they're getting better, as the need for interdisciplinary knowledge becomes more apparent).

They need the machine learning types, people who understand both fields. Heck, most biologists aren't even very good at chemistry. If you start talking to them about nonequilibrium thermodynamics (which is all of life) their eyes glaze over.

Textbook education is one thing, but learning by doing is entirely another.

It's the doing that matters. Science is doing. It's experimentation and publishing and dialoguing with your peers.

 

[Old Rocks]

The exercise with marbles is called Bayesian statistics.

The issue is, how much information did you have before the change, and how much information do you have after the change?

(Question for extra credit: is it possible to lose information from a change?)

If there is a substitution, some of the probabilities change.

If there is a duplication mutation, all the probabilities change all at once.

There are machine learning algorithms that handle these cases. They work on maximizing and minimizing probabilities.

Much of the protein folding AI used in cancer research and drug design works this way.

It is impossible to discuss evolution in scientific terms, without being aware of these mechanisms.

Biologists are generally not very good with math (although they're getting better, as the need for interdisciplinary knowledge becomes more apparent).

They need the machine learning types, people who understand both fields. Heck, most biologists aren't even very good at chemistry. If you start talking to them about nonequilibrium thermodynamics (which is all of life) their eyes glaze over.

Textbook education is one thing, but learning by doing is entirely another.

It's the doing that matters. Science is doing. It's experimentation and publishing and dialoguing with your peers.

On a subject like evolution, you have to involve all science. Because life is a very complex subject, geology, chemistry, physics, all involved. And to understand any of those sciences, much mathematics is necessary. No one person can possibly in a single long lifetime understand all the science involved in life. So you see people who understand almost nothing of any of the sciences denying the existence of evolution, in spite of the evidence of it being present in every cell in their bodies.

 

[Robert W]

Pretty sure this should go in the religion forum.

What religion? Actually this is science.

 

[Robert W]

I have a university textbook on phrenology.

So what?

Do you believe everything you read on Google, or can you think for yourself?

You have an aim, a goal. I don't necessarily believe things on this forum.

 

[scruffy]

You have an aim, a goal.

I study brains. So I have to educate myself in a lot of different areas. And, I try to keep it fun, to avoid getting bored.

Lately I'm into machine learning, to survey the methods they use to get results. Some of them are write-offs from a biology standpoint, but some are interesting.

One of the interesting ones is "variational inference". What's interesting about it is it converts an impossible statistical problem into a difficult but do-able optimization problem. It ends up with a term I call "all possible paths", just like Feynman's physics.

This, it turns out, has a lot to do with life. But very very primitive life, the kind you might find at the quantum level. This is where you might find some of the things I've been talking about, like recursions and such. It takes a minute to wrap your mind around why this is actually life, you have to stare at it for a while but eventually it'll leap out at you.

I don't necessarily believe things on this forum.

Neither do I. But once in a while someone comes up with a gem.

A gem is usually a seed, something I need to know about but haven't yet considered. There's lots of that stuff, "the more I know the more I realize how little I know".

And, my favorite method is to learn by doing. Just like school, first learn the theory then do the lab. And I can't afford to run out and buy an electron microscope every time I get a bug up my butt, so I look for "virtual equivalents" online. The cloud is the obvious answer for computing, but for example I found a thing called "webots" where you can learn robotics in detail without ever touching a wire. So far I've learned all about quaternions, which has to do with everything from forward and inverse kinematics to information geometry. It's an invaluable tool for brain modeling. But the point is I can get online and play with robots and have fun, and actually learn something useful by doing it.

So this "all possible paths" thing is very cool. How do you calculate that, if you don't know what the paths are? Well, it's the same thing as "missing information", like trying to solve the marbles problem without knowing how many marbles are in each bag. Turns out, there's a way to do it, that's almost like curve fitting. You optimize around the KL-divergence and drive it to zero. In higher dimensions (more unknowns) you can do it recursively, using a very simple algorithm - and our brains can do it in parallel in real time.

Well, if you distill the essential math, you discover that an individual cell can do it too, at the biochemical level - because all it really is, is "compartments" of information. In the brain you let the synaptic weights equilibrate, and in a cell you let the chemical reactions equilibrate. Either way, you "estimate" what "all possible paths" means. In the brain you use it to find a connection between seemingly unrelated information, and in a cell you use it to find an evolutionary pathway. Cells are very smart, they can do stuff like that. Our immune systems are examples, we find a solution for how to control an unknown infection that's never been seen before. The cell surface receptors will pick up the antigen, but somehow we have to get from there to "which genes do I need to amplify", starting with an estimate of all possible paths.

See what I mean? Fossils have nothing to do with, it's the "how do I" that matters.

 

[JoeB131]

You get to walk slowly through this scientists discussions where his aim is to prove evolution is an adult fairy tale.

Be careful since he just may persuade you that evolution as taught is an adult fairy tale.

Well, first, he's not talking about evolution, he's talking about abiogenesis (how chemical reactions caused the first life to happen)

And I'll freely admit, we don't know everything about that, yet.

Evolution is something totally different.

It holds that life changes over time through adaptive radiation and natural selection.

That is not only provable through the fossil record, but we can see it in real time as viruses develop immunity to medicines and vaccines, causing us to create new medicines and vaccines. Or how insects develop an immunity towards certain insecticides.

 

[Robert W]

I study brains. So I have to educate myself in a lot of different areas. And, I try to keep it fun, to avoid getting bored.

Lately I'm into machine learning, to survey the methods they use to get results. Some of them are write-offs from a biology standpoint, but some are interesting.

One of the interesting ones is "variational inference". What's interesting about it is it converts an impossible statistical problem into a difficult but do-able optimization problem. It ends up with a term I call "all possible paths", just like Feynman's physics.

This, it turns out, has a lot to do with life. But very very primitive life, the kind you might find at the quantum level. This is where you might find some of the things I've been talking about, like recursions and such. It takes a minute to wrap your mind around why this is actually life, you have to stare at it for a while but eventually it'll leap out at you.



Neither do I. But once in a while someone comes up with a gem.

A gem is usually a seed, something I need to know about but haven't yet considered. There's lots of that stuff, "the more I know the more I realize how little I know".

And, my favorite method is to learn by doing. Just like school, first learn the theory then do the lab. And I can't afford to run out and buy an electron microscope every time I get a bug up my butt, so I look for "virtual equivalents" online. The cloud is the obvious answer for computing, but for example I found a thing called "webots" where you can learn robotics in detail without ever touching a wire. So far I've learned all about quaternions, which has to do with everything from forward and inverse kinematics to information geometry. It's an invaluable tool for brain modeling. But the point is I can get online and play with robots and have fun, and actually learn something useful by doing it.

So this "all possible paths" thing is very cool. How do you calculate that, if you don't know what the paths are? Well, it's the same thing as "missing information", like trying to solve the marbles problem without knowing how many marbles are in each bag. Turns out, there's a way to do it, that's almost like curve fitting. You optimize around the KL-divergence and drive it to zero. In higher dimensions (more unknowns) you can do it recursively, using a very simple algorithm - and our brains can do it in parallel in real time.

Well, if you distill the essential math, you discover that an individual cell can do it too, at the biochemical level - because all it really is, is "compartments" of information. In the brain you let the synaptic weights equilibrate, and in a cell you let the chemical reactions equilibrate. Either way, you "estimate" what "all possible paths" means. In the brain you use it to find a connection between seemingly unrelated information, and in a cell you use it to find an evolutionary pathway. Cells are very smart, they can do stuff like that. Our immune systems are examples, we find a solution for how to control an unknown infection that's never been seen before. The cell surface receptors will pick up the antigen, but somehow we have to get from there to "which genes do I need to amplify", starting with an estimate of all possible paths.

See what I mean? Fossils have nothing to do with, it's the "how do I" that matters.

I understand vocabulary and I understand experts. I will never pretend to have your vocabulary nor your expertise.
To give you an example, you mentioned Recursively!!! I opened my university book about evolution and that word is not even used in the book.

I expect I can do the same thing to you in my own fields of expertise.

 

[task0778]

I see no reason why the concept of evolution and God have to be diametrically opposed to each other. God may have started life on this planet and then used evolution to change it from the simplest one-celled organisms to the abundant life forms that exist today. I really don't see the point in arguing about it since there's no way to prove anything anyway.

 

[Robert W]

That is not only provable through the fossil record, but we can see it in real time as viruses develop immunity to medicines and vaccines, causing us to create new medicines and vaccines. Or how insects develop an immunity towards certain insecticides.

Thank you JoeB131. Joe, Virus are not living whatchamacallits.
Fossils are not living.
Abiogenesis is not evolution?
Change from non life to living is not evolution?
Immunity is evolution?

 

[Robert W]

I see no reason why the concept of evolution and God have to be diametrically opposed to each other. God may have started life on this planet and then used evolution to change it from the simplest one-celled organisms to the abundant life forms that exist today. I really don't see the point in arguing about it since there's no way to prove anything anyway.

They don't have to be opposed. I started this to expose posters to a particular scientists views for purposes of education.

 

[scruffy]

I understand vocabulary and I understand experts. I will never pretend to have your vocabulary nor your expertise.
To give you an example, you mentioned Recursively!!! I opened my university book about evolution and that word is not even used in the book.

Because the university texts think evolution is about fossils and bags of water.

Does your book talk about chemical evolution?

How about stable atoms evolving from quantum fluctuations, does it talk about that?

How about a human being evolving from an embryo?

The humans in the Tibetan mountains are evolving as we speak, we can see it in action. Their incidence of hypoxia decreases every year.

An interesting case study is Neisseria gonorrhoeae, the monococcus that causes gonorrhea. It's polyploid, it has more than one copy of its own DNA. And, it can endocytose DNA from the environment and from other bacteria, including other Neisseria species.

You've heard of antibiotic resistance, well, in Neisseria this happens at least 5 different ways. Mutations in efflux pumps, mutations in antibiotic binding proteins, mutations in RNA polymerase, etc etc. Whatever can mutate, does. And because of the polyploidy, one mutation is sufficient.

Here's a case study where six different kinds of resistance evolved from exposure to a single antibiotic. The evolution occurred within days. Generation time of Neisseria is 60 minutes, times 138 cultures. About 5 days.

In vitro selection of Neisseria gonorrhoeae unveils novel mutations associated with extended-spectrum cephalosporin resistance - PMC

The emergence of Neisseria gonorrhoeae strains resistant to extended-spectrum cephalosporins (ESCs) is a worldwide concern because this class of antibiotics represents the last empirical treatment option for gonorrhea. The abusive use of ...

pmc.ncbi.nlm.nih.gov

I expect I can do the same thing to you in my own fields of expertise.

Maybe. You can try.

 

[Robert W]

Because the university texts think evolution is about fossils and bags of water.

Does your book talk about chemical evolution?

How about stable atoms evolving from quantum fluctuations, does it talk about that?

How about a human being evolving from an embryo?

The humans in the Tibetan mountains are evolving as we speak, we can see it in action. Their incidence of hypoxia decreases every year.

An interesting case study is Neisseria gonorrhoeae, the monococcus that causes gonorrhea. It's polyploid, it has more than one copy of its own DNA. And, it can endocytose DNA from the environment and from other bacteria, including other Neisseria species.

You've heard of antibiotic resistance, well, in Neisseria this happens at least 5 different ways. Mutations in efflux pumps, mutations in antibiotic binding proteins, mutations in RNA polymerase, etc etc. Whatever can mutate, does. And because of the polyploidy, one mutation is sufficient.

Here's a case study where six different kinds of resistance evolved from exposure to a single antibiotic. The evolution occurred within days. Generation time of Neisseria is 60 minutes, times 138 cultures. About 5 days.

In vitro selection of Neisseria gonorrhoeae unveils novel mutations associated with extended-spectrum cephalosporin resistance - PMC

The emergence of Neisseria gonorrhoeae strains resistant to extended-spectrum cephalosporins (ESCs) is a worldwide concern because this class of antibiotics represents the last empirical treatment option for gonorrhea. The abusive use of ...

pmc.ncbi.nlm.nih.gov

Maybe. You can try.

Checked my university book by Mark Ridley, Evolution.
Nothing in it about water.

 

[JoeB131]

Thank you JoeB131. Joe, Virus are not living whatchamacallits.
Fossils are not living.
Abiogenesis is not evolution?
Change from non life to living is not evolution?
Immunity is evolution?

It's not just viruses, it's bacteria, insects, etc. that we see grow immunities.

Abiogenesis is not evolution.

That you are ignorant of terms and what they mean is the problem.

That and you think that a book of bronze age fairy tales explains much of anything.

 

[scruffy]

Checked my university book by Mark Ridley, Evolution.
Nothing in it about water.

Not even that?

Doesn't sound like a very good text.

No discussion of hydrogen sulfide? Copper catalyst? Micelles and lipid bilayers?

Mark Ridley was educated in the 80's. That's 40 years ago.

Try this on for size:

https://www.amazon.com/Fortunate-Events-Chance-Making-Planet-ebook/dp/B086YHXHPL?tag=usmb-20

Make sure you click on the down arrow to see Carroll's other books.

And this is the serious no-bullshit text on genetics:

https://www.amazon.com/Evolution-Douglas-J-Futuyma/dp/1605356050?tag=usmb-20

 

[scruffy]

To understand evolution, start at the very beginning.

Earth, 4.6 billion years ago.

Temperature: 3600 degrees F
Atmosphere: None

Then, 4.3 billion years ago

Magma oceans covering most of Earth's surface
Volcanic degassing creates a sparse atmosphere of H2O, CO2, NH3, CH4, and other stuff like neon and xenon.
As magma degrades it leaves behind large quantities of iron and sulfur.

Then, 3.8 billion years ago

Intense meteorite impacts, known as the Late Heavy Bombardment Period.
Chondritic material deposited all over the earth

Then,

Earth begins to cool
Early ocean formation
Ocean temperatures 130-180 degrees F

Compare this with the estimated age of the earliest cyanobacter, which is about 3.5 billion years.

Cellular life started "immediately" upon Earth's cooling, meaning the chemical precursors were already present.

You had half a billion years of chemical evolution before biological evolution started.

 

[Robert W]

Amazon.com

I bought this book due to your recommendation.