"I'm treating the mutation rate as a substitution rate" - Dr. Nathaniel Jeanson

A tree is a tool of the scientific method for many applications. Fitting the data to a tree does not however isolate cause.

A tree pattern simply needs to fit a tree better then the null or a random pattern. There are many possibilities depending on the exceptions.

I am asking specifically how gene loss and gain can account for the Howe including a model of how the change can result in the observed gene pattern.

This is with 6 lottery balls. What happens if you increase the number of balls to 30 assuming 50 different numbers on each lottery ball. How often will someone win the lottery?

The quantifiable tree-like signal in the data is a test of the proposed causes which are modification of traits combined with vertical inheritance. This has been the test since the time of Darwin.

Since the mid 1800’s we have also discovered the genetic underpinnings of how the modification works which is mutation, natural selection, drift, and vertical inheritance.

You have been shown all of these causes before, and yet again you act like you have never seen them. Why?

Just as there are many patterns of data points that will produce a statistically significant positive correlation:

That doesn’t change the fact that there is a correlation. You are failing to understand the basics of statistics.

You have been shown this multiple times.

I asked you to explain this scenario:

Let’s use the lottery as an example. There is a single winner of the lottery. You calculate the odds of the winner you observed, and the odds are 1 in 150 million. And yet, it just took 1 drawing for that person to win. How do you explain this?

Can you do it or not?

2 Likes

I haven’t forgotten. Got sick and then was thinking a lot about the other thread, plus holiday weekend. Still want to go through the paper you shared (and have shared before; kept getting distracted instead of reading it).

Okay fair enough. Take the time you need.

1 Like

The model is simple. Genes are gained, at a rate we can estimate from the data. Genes are lost, at a rate we can estimate from the data. Gain is by duplication of an existing protein-coding sequence or by change to a non-protein-coding sequence or by recombination among sequences. Loss is by deletion or inactivation (early termination, loss of promoter, etc.). It should be noted that the tree has vanishingly few patterns for which two events are required; almost every pattern requires only a single gain or loss. The latter are, in other words, perfectly compatible with the standard tree but not with other trees. Why should that be?

3 Likes

Hi, Bill. Just so we’re all on the same page, can you give a quick summary of the responses you’ve received on this question? Thanks!

1 Like

This is not a population genetics model. The reality is there is no way to reconcile the data mathematically without a deterministic mechanism driving the change. I think it is safe to eliminate random gene change along with selection and drift as the cause.

John, I think this pattern is clearly beyond the edge of currently understood evolutionary mechanisms. Common descent is at best a partial explanation without a mechanism you can create a population genetics model with.

Answer the question I asked and you will see why your scenario fails. The number of arrangements increases exponentially with the length of the sequence. Lotteries have short sequences of 5 to 6 elements…

T with all due respect your answers lack substance. Making a statistical argument without a p value or statistical estimate for the probability of your claim is meaningless. Asserting mechanisms like selection and drift without demonstrating they account for the pattern is meaningless.

The problem here is no mechanism for the pattern has been identified. When or if a mechanism is identified there will not be a problem building a population genetics model showing the origin of the pattern…

What happens if you increase the number of balls to 30 out of 50? People who understand probability will know that you don’t.

Hint: The probability of each ticket winning the lottery decreases as you increase the number of balls that need to be matched only until you get to 25 balls. Then the probability of winning starts to increase again.

OK, new rule … Bill is not allowed to reference probabilities and p-values without first consulting a statistician.

My fees are very reasonable. :grin:

5 Likes

Hi Matt
I have received no substantive response to this question as it requires an identified mechanism to explain the pattern. So far the mechanism has not been identified. If the mechanism is not powerful enough a population genetics model is not possible.

More word salad. Reconcile the data mathematically with what? Why a deterministic mechanism rather than a stochastic one? Why is it safe to eliminate any of those factors?

Do you have any actual basis for your opinion? So far you have shown nothing.

As has been explained to you many, many times, common descent explains the pattern of the changes, not the changes themselves. The fit to a tree shows two things: that the tree is real and that the changes actually happened. How or why they happened is another question, but that question doesn’t need to be answered in order to accept the tree. You are just making excuses to ignore the data.

3 Likes

Once more you confuse the pattern — nested hierarchy — with the elements of that pattern. Population genetics doesn’t explain the pattern; a tree explains the pattern.

And yet you also feel able to say this:

Wasn’t that a statistical argument, or at least a statistical claim? Where are your p values?

2 Likes

To me, it is self evident that a bunch of genes will be shared as inherited from a common ancestor, and distinct populations will each develop independent and unique genes. Given that it is not disputed that random gene change along with selection and drift, actually happens; what do you think the outcome of that process will look like in terms of genetic phylogeny and Venn type diagrams? How could the pattern be any different?

3 Likes

Because you cannot generate a population genetic model with a stochastic mechanism and reconcile the pattern so it is safe to eliminate stochastic mechanisms for all hypothesis.

Yes.

The “changes” are not established at this point. You are assuming the differences are due to changes and not separate starting points.

Hi, Bill. I understand you don’t find the responses substantive, but could you at least summarize what they are saying? Thanks!

3 Likes

Hi Ron
The question is if the magnitude of the change observed is in reach of stochastic mechanisms. If you look at either Lynch’s proposed population genetics model or Behe’s the answer is no.

???
Bill, I think you might not understand the meaning of “tree”.

3 Likes

Hi Matt
The counter claim is the pattern is explained by common descent plus gene gain and loss which is accumulated in the population by selection or genetic drift.

2 Likes

So the pattern is no problem, it is just a matter of degree?