Natural Selection or Luck

[notto]

If Chance of survival without mutation = 20% then
Improved chance of survival for that remaining 20% = 0.1%.

Again, you need to look at large populations of animals to do these things. If out of 100,000, only 20% survive to breeding age, then there is still 20,000 who can can be acted on by selective presure where the mutation can be beneficial.

Of course with this hypothetical survival rate of 1 out of 5, selective pressure would likely act on any mutation that causes the breeding mothers to have more offspring. Large litters means more genes from that litter survive, even with high mortality rates.

 

[lucaspa]

Today at 04:11 AM Micaiah said this in Post #56

These discussions make it clear that when considering a single beneficial mutation in an individual, the percent increase in the chances of survival is low. Lucaspa has suggested 0.001 or 0.1% as a posssible increase in the chance of survival of a single point mutation.

Micaiah, you are not reading what I say.  Listen carefully.  The percent increase in survival can range from negative to 100% depending on the mutation and the environment.  If the selection pressure is severe enough, then everyone dies except the individual with the mutation.  That is what happened with antibiotic resistance.   

What I am saying is, no matter how small the increase in the chance of survival, natural selection will ensure that the mutation spreads thru the entire population.

If enough of these are cobbled together, then evolutionists say we end up with the change that is postulated.

Not postulated, observed.  And it's not an "if".  It's mathematically certain.  Because the change is cumulative.  The tuskless elephant is a developmental macromutation, but even a mutation with a small effect -- a slightly bigger hoof in horses -- is going to be fixed.  Then when a new mutation comes along for an even bigger hoof, it will build on the bigger hoof that is already there.  The process is inevitable.

In the case of the horse discussed above, it was the change from for digits to a single hoof. The chances of the required number of mutations occuring in the required order is very low.

Natural selection, because it is an algorithm, lowers the odds, because of the cumulative effect. 

I consider the chances are so low, and the survival benefit gained from each step is small enough that we can say the changes are essentailly spontaneous and random.

The changes are indeed spontaneous and random with respect to the needs of the individual and population.  That is variation.  But you are forgetting that natural selection is a TWO-STEP PROCESS. 

1. Variation.
2. Selection.

While the first step does appear to be random -- which is what causes troubles for creastionists -- the second step is pure determinism.  That's what the equations show.

Here is a simplified example of how cumulative selection can cut down "odds"

You have a 1 in 1024 chance of correctly winning 10 coin tosses in a row.  But I can guarantee you I can find someone who can do so.  How?  Simple, use cumulative selection in the form of a single elimination tournament.  I start with 1024 people and pair them up.  Then each pair tosses a coin.  The 512 winners are selected to go to the next round.  Again they are paired and do a coin toss; the 256 winners are selected to go to the next round.  Repeat this 7 more times.  Now you have 2 people who have won 9 coin tosses in a row.  The winner of this round has won 10 coin tosses in a row.  And it is a certainty that such a person will be found with this method.  We have taken odds of 1 in 1024 and converted that into virtual certainty.  Now, I don't know which individual will win the tosses, but it is certain that one of them will, given the algorithm of the competition.  Evolution by natural selection is a competition algorithm, more complex but analogous to the single elimination tournament algorithm.
[/QUOTE]

 

[lucaspa]

Today at 04:25 AM Micaiah said this in Post #58

Lucaspa, you must accept that many deaths occur mainly as accidents.
Even if you don't , you should recognise the causes may often be unrelated to the development of the hoof. Some of these are evident in the examples I gave above. Another would be diseases. The beneficial hoof mutation would have little impact on the horses chance of surviving in these cases. 

What appears to be "accident" to you is selection at work.  What is "accident" for the individual is really determinism when looking at a lot of individuals.  Take a look at the start of a triathalon when all the swimmers hit the water at the same time.  A lot of kicking and shoving.  What appears to be an accident for one swimmer -- getting kicked and being slowed down -- is actually a selection process picking those swimmers that can swim faster despite being kicked and shoved.

What you are trying to do is separate all the different selection pressures.  But that doesn't happen.  Individuals are born with larger and smaller hooves from the mean and with greater and lesser resistance to disease from the mean. Obviously there are going to be individuals with both an increased hoof size and an increased resistance to disease. So guess what happens? Only the individual with both is going to make it.  Those with smaller hoofs but resistance to disease will be eliminated by the jump and those with larger hooves and lesser resistance to disease will be eliminated by the illness.

We are talking about two things here. One is the chance of the animal surviving, and the other is the gain in the animals chance of survival as a reult of a single mutation, such as could cause a change to the hoof.

No, we are talking about the same thing here.  The chance of the animal's survival depends on the genes (traits) it has.  If the animal has a larger hoof in an environment where larger hoofs are an advantage, then the animal's chance of survival is increased.

 

[notto]

When you play the lottery, the more people who enter does not affect your individual propability of winning but if you look at the population that enters, the larger the population, the better the probability of someone winning increases.

 

[lucaspa]

Today at 07:28 AM Micaiah said this in Post #60

What is the chance of an animal with a mutation surviving. What is the chance of this mutation surviving in future generations. How do you calculate these probabilities.

In this case, we've used the following parameters:
- Chance of survival without mutation - 20%
- Inproved chance of survival with mutation - 0.1%

Micaiah, you've used the parameters.  We've said the parameters are invalid. 

So, let's correct your scenario.

What you are saying is that you have a population where each pair has an average of 5 offspring. Since the population is stable, only 40% of the offspring survive to reproduce.  What matters is which of those individuals survive.  As I pointed out, that doesn't mean that 2 offspring of each pair survive, but just that 40% do.

Putting numbers on this, you have 5,000 breeding pairs producing 25,000 offspring each generation.  Of these only 10,000 survive to reproduce.  You are thinking that is is "accidents" that keep a population stable. But that isn't true.  True accidents won't do it.  Look at the increase in human population in the last several generations as we removed most of selection.  Our population isn't stable because true "accidents" aren't enough to keep it stable.

So, in your example, all those "accidents" are selection. If they didn't exist (predators, disease, etc) you would have 25,000 in hte next generation and 60,000 in the generation after that, and 150,000 in the next generation.  It is selection, not accidents, that keep the population in check.

So, let's start out with a mutation that increases fitness by 50% = 0.5

p equals the proportion of individuals with the mutation and q equals the proportion with the old mutation.  Starting off, we have p = 1/10,000 and q = 9,999/10,000.  Doing the equation we find that delta p = 0.25.  That means the frequency increased from 1 in 10,000 to 1 in 4.  Of the 10,000 who survive to adulthood that means that 2,500 will now have the mutation.  In one generation.

Now, let's take a less extreme, where w = 0.01 or 1%.  Now delta p is only 0.005 but that means that now 5 individuals will have the mutation in that generation instead of 1.

Obviously, if you keep this up over generations, eventually everyone will have the mutation.

Now that I look again at what you are saying, the answer is that the mutation increased the chances of survival of the individual that carries it by the fitness value of the mutation. 

I think what you are trying to do is say what the fitness value is before selection.  In practice, that is not done because we can't.  We can't know all the variables and selection pressures a population faces.  What we do know from Mendelian genetics is that, if there are no selection, then p would always be 1 in 10,000.  But then, if there were no selection, population would go up to 25,000 in the first generation (no selection to keep the population stable), then up to 60,000 the next generation, etc.

 

[Cantuar]

Lucaspa, you must accept that many deaths occur mainly as accidents.
Even if you don't , you should recognise the causes may often be unrelated to the development of the hoof. Some of these are evident in the examples I gave above. Another would be diseases. The beneficial hoof mutation would have little impact on the horses chance of surviving in these cases.

You're ignoring the fact that these neutral or beneficial mutations spread thoughout populations. So even if a particular individual did happen to meet with an accidental death, it wouldn't make a difference because you have a whole population with that mutation (or at least with that mutation in a significant fraction of members), and it's massively unlikely that every member of the population with a particular mutation will meet with an accidental death when the ones that don't have that mutations don't die accidentally. If such an extremely unlikely thing did occur, you'd be looking at a case where the mutation was probably involved in the deaths, in which case they wouldn't be accidental.

 

[lucaspa]

Today at 02:35 PM Cantuar said this in Post #66



You're ignoring the fact that these neutral or beneficial mutations spread thoughout populations. So even if a particular individual did happen to meet with an accidental death, it wouldn't make a difference because you have a whole population with that mutation (or at least with that mutation in a significant fraction of members), and it's massively unlikely that every member of the population with a particular mutation will meet with an accidental death when the ones that don't have that mutations don't die accidentally. If such an extremely unlikely thing did occur, you'd be looking at a case where the mutation was probably involved in the deaths, in which case they wouldn't be accidental.

Micaiah is trying to look at the instance where the mutation first appears and is in only one individual.  Where he goes wrong is in claiming as "accidents" what is, in reality, selection.

The colt that wanders too far from Mom and is eaten by a predator is selected.  The colts with genes that kept them closer to Mom have an advantageous variation. Predation is selection, not accident. So is disease. So is the ability to avoid a fall. All  Micaiah's examples are selection and not accident.

Now, it is possible that, while a mutation for a larger hoof is only in one individual, then one of the other selection pressures will kill it.  However, Micaiah's second strawman is that variations/mutations to get a larger hoof are going to be rare.  That is, only one animal in the history of horses will have a mutation to get a larger hoof.  So, if that one is killed by another cause, then larger hoof is lost forever.

However, it is documented that there is enormous variation in any population of animals. Therefore in any sizeable population (over 50) it is virtual certainty that there will be at least 2 horses with hoof size greater than the mean -- giving a selective advantage in that environment.  This is what you are getting at in your post.  Also, even if that one animal in that generation were killed, the hoof doesn't shrink.  Instead, you have to simply wait until the next generation or the next.  Among all the random variations, there will be another one for a larger hoof.  Not the same, perhaps, as the first, but one that will work.

And that last is part of the contingency of evolution that creationists hate because it smacks of "undirected".

 

[Micaiah]

We are not talking about the increase in hoof size. We are talking about a hoof evolving from a foot with four digits. What would the chances be of a single point mutation occuring? How many animals could we expect in a population at any one time, or if they are rare, over a given period of time?

Note that the chances of this occuring are different to the chances of the animal surviving.

 

[Cantuar]

Micaiah is trying to look at the instance where the mutation first appears and is in only one individual._ Where he goes wrong is in claiming as "accidents" what is, in reality, selection.

Yes, I understand that, but on the other hand I think it's a bit unrealistic to say that there are NO accidental or random deaths in the wild (if that's what you were saying). Obviously Micaiah is stretching when he attributes all those deaths to accident when many of them are clearly selection based, but if a herd is running along under a cliff when some rocks start falling, there's bound to be an accidental component to the outcome in terms of who survives and who doesn't.

 

[Micaiah]

The deaths often result from causes that are entirely unrelated to the mutation. In this population I'm saying 80% die, and of the 20% that live, the increase in their chance of survival is largely unaffected by the mutation. If we use 0.1% as the increase in the chance of survival, then the chance of survival becomes 20.02%. Obviously the benefit from the mutation is minimal.

The mutation is entirely random. The chance of survival is practically independent of the mutation and is therefore random with respect to the mutation.

In his example above, Lucaspa discusses the inevitability of getting a number of outcomes in a row from flipping coins. In likening this example to mutations in the animal population, he assumes that there can be a number the same mutations in a population at once. The chances of a mutation replicating itself would be greatly increased if that were the case, but you need to show this scenario is realistic. My understanding of point mutations is that they result from copying errors in the formation of new cells or gametes. The rate of those copying errors is very low. The chance of getting an error in one specific nucleotide is very low. The chance of getting the error with the required base is even lower (divide by 3 since there are three bases). Once you have the required mutaion, you have to repeat it a number of times.

Here is another example to illustrate. If you had say 3 coins on the floor in a straight line each separated by a distance of 200mm (your turn to convert units), most people would assume that they were positioned in that order by someone (Intelligent design). The chance of getting the coins to line after tossing them in the air would be very unlikely. The chance of getting three mutations in a specific nucleotide in unison is like the example of the coins.

 

[chickenman]

micaiah, evolution is much more complex than your analogies

if you want to see the affects of selection coefficients on allele fixation, then download something like visual popgen

favourable alleles are swept to fixation depending on their degree of dominance and their selection coefficient - a statistical program like popgen will show you

 

[notto]

Today at 03:34 AM Micaiah said this in Post #70

Here is another example to illustrate. If you had say 3 coins on the floor in a straight line each separated by a distance of 200mm (your turn to convert units), most people would assume that they were positioned in that order by someone (Intelligent design). The chance of getting the coins to line after tossing them in the air would be very unlikely. The chance of getting three mutations in a specific nucleotide in unison is like the example of the coins.

Random Mutation throws 1000 coins. Selective pressure discards all except the 3 in a line. No intelligence needed.

Or, to look at your example another way, if I take three coins and throw them 1000 times, there is a chance that they would land separated a distance of 200mm. If I throw them a million times, the chance is higher. Of course, there is also a chance that I get it the first time. Again, no intelligence necessary, just selective pressure (what you are looking for in the results) to select when it is right.

Your example can be used to show how selective pressure can build on favorable results. You need to know what number of unsuccessful attempts were made to try to determine the probability of the outcome. Otherwise, any argument from probability fails. In your example, you are using only one attempt. This is not analgous to random mutation within a population.

 

[lucaspa]

20th March 2003 at 11:53 PM Micaiah said this in Post #68

We are not talking about the increase in hoof size. We are talking about a hoof evolving from a foot with four digits. What would the chances be of a single point mutation occuring? How many animals could we expect in a population at any one time, or if they are rare, over a given period of time?

Note that the chances of this occuring are different to the chances of the animal surviving.

A "hoof" is simply an enlarged toe -- the middle one.  So, when I say "larger hoof" I am saying "increased size of the middle toe".

The average mutation rate in eukaryotes is about 1 per 10^6 to 1 per !0^9 nucleotides.  In other words, about 1 in a million to 1 in a billion nucleotides.  The human genome is 6 billion nucleotides, so that is 6 to 6,000 mutations per individual. 

Calculating the number of individuals with a mutation in a particular gene is a bit trickier because some parts of the genome are very resistant to mutation while other parts are very susceptible to mutation.  It also depends on the population size. Finally, it depends on how many genes contribute to the size of the toe.  We are assuming that it is one, but most traits like that are polygenic -- several genes contribute. So a mutation in any one of them would increase toe size.

Take your population of 10,000 individuals. Assume 30,000 genes as in humans and 6 to 6,000 mutations per individual.  That's 60,000 to 60 million mutations in the population.  If they were equally distributed over the genes, that would be 2 to 2,000 individuals with mutations in the gene controlling toe size per generation.  Those numbers go up if there are more than one gene involved.

The "chance of surviving" is non-random depending on the design of the animal.

 

[lucaspa]

Today at 04:34 AM Micaiah said this in Post #70

The deaths often result from causes that are entirely unrelated to the mutation.

That's the mistake.  No, they don't.  The deaths are related to the design of the animal, which in turn is dependent on the form of the gene (allele) the animal has. 

In this population I'm saying 80% die, and of the 20% that live, the increase in their chance of survival is largely unaffected by the mutation.

That's the mistake.  That 80% die is due to the selection of individuals with the best allele in the population. 

If we use 0.1% as the increase in the chance of survival, then the chance of survival becomes 20.02%. Obviously the benefit from the mutation is minimal.

Even if your calculation were correct, the equations show that there is no such thing as "minimal" when it comes to selection picking the mutation that increases fitness.  No matter how small, 0.1% or 0.000001%, as long as the mutation increases fitness, eventually it will spread through the entire population.  It will take more generations, but the outcome is assured.

The mutation is entirely random. The chance of survival is practically independent of the mutation and is therefore random with respect to the mutation.

You are trying to separate the two-step process that is natural selection.  That is called making a strawman.  The mutation is random with respect to the needs of the individual or the population.  That is, a mutation is just as likely to give a smaller toe as a larger one.  However, the chance of survival is tied to the size of the toe, and therefore tied to the mutation.  The individuals with the mutation for a larger toe will have increased fitness and therefore an increased chance of survival.  Yes, the selection pressure is always there, but that doesn't make it "independent" of the mutation as you are using the term.

In his example above, Lucaspa discusses the inevitability of getting a number of outcomes in a row from flipping coins.

No, I'm discussing the inevitibility of getting a person with those outcomes.  Different thing.  Read it again carefully, Micaiah.

In likening this example to mutations in the animal population, he assumes that there can be a number the same mutations in a population at once.

I'm not assuming. Go back up a number of posts and look where three mutations with the same effect were present in a population at once.  That's not an assumption, but a conclusion from data.  The same thing is documented in population genetics studies where it is documented that individuals with the same traits are present in the population at the same time: larger and smaller middle toes from the mean. 

Here is another example to illustrate. If you had say 3 coins on the floor in a straight line each separated by a distance of 200mm (your turn to convert units), most people would assume that they were positioned in that order by someone (Intelligent design). The chance of getting the coins to line after tossing them in the air would be very unlikely. The chance of getting three mutations in a specific nucleotide in unison is like the example of the coins.

The fallacy here is that there are several changes in a gene that will end up with the same macro result: larger toe size.  You are making the same GIGO mistake all creationists make in the probability calculation: there is one and only one change that will have the effect. That has been documented again and again to be untrue.  Again, go back to the post earlier in the thread with the abstract on mutations in Bactillus.  You didn't read it or didn't understand it. 

 

[Micaiah]

Are you suggesting that we simply need to change one nucleotide to develop a hoof from a four digit foot. If that was the case, I'd expect to see a lot of foxes running around with partially formed hooves.

Even if it were possible to obtain the same physical changes from genetic changes, this wouldn't improve the chances of a single point mutation spreading to the whole population which is what evolution assumes. For that to occur, we need a number of breeding animals with an identical mutation ie. the same point mutation.

The question is "What are the chances of the same point mutation occuring a number of times?"

 

[DNAunion]

Lucaspa: Even if your calculation were correct, the equations show that there is no such thing as "minimal" when it comes to selection picking the mutation that increases fitness. No matter how small, 0.1% or 0.000001%, as long as the mutation increases fitness, eventually it will spread through the entire population. It will take more generations, but the outcome is assured.

DNAunion: WRONG!!! Beneficial mutations are often times lost from a population - those that do never become fixed.

Lucaspa, do you know ANYTHING about biology?

 

[chickenman]

DNAunion vs Lucaspa continues

you're so petty

lucaspa is correct though, given enough generations, the mutation-selection balance will cause beneficial mutations to become fixed.
Even if they're lost on the first occaision they occur, each subsequent appearance of the mutation represents another chance for it to become fixed

and mutations destined to become fixed are fixed quickly

 

[lucaspa]

Yesterday at 10:19 PM DNAunion said this in Post #76



DNAunion: WRONG!!! Beneficial mutations are often times lost from a population - those that do never become fixed.

Look at the equation, DNAunion.  Even the slightest positive selection value will result in fixation.

Now, since you say that loss of beneficial mutations from a population is a common event, you should give us examples of this happening.

I suspect you are discussing a different phenomenon, but let's see what you say.

 

[lucaspa]

Yesterday at 10:19 PM DNAunion said this in Post #76



DNAunion: WRONG!!! Beneficial mutations are often times lost from a population - those that do never become fixed.

Lucaspa, do you know ANYTHING about biology?

Is it constitutively impossible for you to hold a discussion without insulting the other person?

 

[lucaspa]

Yesterday at 09:48 PM Micaiah said this in Post #75

Are you suggesting that we simply need to change one nucleotide to develop a hoof from a four digit foot.

No, although changing one nucleotide does convert a multilegged organism to one of six legs. More on that later.

A horse's hoof is a modified nail on the middle toe. The other toes have retracted to the point where they either don't exist or are just very small bones.  What I'm saying is that there are several different mutations that will result in each step of the modification of the nail of the middle toe to a hoof.  It is not one mutation jump nor just one mutation that can make the nail a little bit larger.  Instead, several mutations will have the same phenotypic effect. (phenotype is what you see on the animal).

For instance, humans make 3 different types of hemoglobin.  A fetal, newborn, and adult forms.  All of them transport oxygen in the blood.  Three different genes, all do the same job (phenotype).  So, each step in changing the size of the nail and the size of the bones in the middle toe can be brought about by several different mutations.

If that was the case, I'd expect to see a lot of foxes running around with partially formed hooves.

Selection would weed out any individual foxes with those types of modifications because a hoof wouldn't fit with the lifestyle of a fox.  Foxes use claws to grasp their prey. A hoof can't grasp. So, any individual fox with a modification toward a hoof would be at a selective disadvantage to other foxes without the modification and that allele would be eliminated from the gene pool.

Even if it were possible to obtain the same physical changes from genetic changes, this wouldn't improve the chances of a single point mutation spreading to the whole population which is what evolution assumes. For that to occur, we need a number of breeding animals with an identical mutation ie. the same point mutation.

The equation shows otherwise.  It is based on calculating the proportion of the allele, which is the mutation. The spreading of a point mutation is not an "assumption", but an observation backed by mathematics.

Now, selection works on the phenotype, not the genotype.  So, if you have several different alleles that all result in the same or similar phenotype, all the mutations will be selected for initially.  However, there will be slightly different fitness values for each mutation because the phenotypes will not be identical, so eventually either 1) there will be a balance of alleles in the population or 2) one allele will replace all the others.

The question is "What are the chances of the same point mutation occuring a number of times?"

That's not the question. Rather, the question is: what is the chance of any mutation or recombination with a favorable trait showing up in a population?  You are mistakenly saying that there is one and only one mutation that can get the job done, but the data shows that is not true. 

That chance of a beneficial variation is not 1 because species do go extinct.  But it is a large fraction of 1.