Is Helicase a House of Cards?

Yes, he certainly does.

Interestingly, the HOGENOM database is limited to microbes, so zebra finch and zebra fish should not be there at all. Perhaps he meant to say HOVERGEN? Now, how many other teleosts and birds are in that data set? How many other vertebrates?

I’m not sure why those two species would be predicted to share unique gene families based on ID. Is it because they both have the word “zebra” in their common names?


IIRC, a criticism/suggestion was that Ewert was relying on datasets many of which were almost certainly incomplete, even fragmented. I think his paper may be better viewed as a framework that might be applied once suitable datasets are available.

That’s my recollection, at least. I may be totally wrong.

1 Like

I see two issues here. The first is your representation of the problem. Let’s set that bit aside for now.

Sal, respectfully, that’s not LLN. It might be a joint probability of some sort. I’m a statistician, and if I’m not looking at it right then it you are not describing what you think you are describing. If you want to describe the distribution of repeating sequences then you might have a binomial or multinomial.

Correction: In a comment above I mentioned Kolmogorov’s Inequality, but I meant Chebyshev’s inequality. I will edit!


Take 500 fair coins and flip the first one and paint the number “1” in WHITE on the face (heads/tails) that is exposed upward. Do the same for the next coin, flip it, paint the number “2” in WHITE on the face (head/tails) that is exposed upward. Do that for all 500 fair coins.

What’s the probability the all the coins will be flipped on the next round with the WHITE side facing up? It’s the binomial distribution all over again with P(WHITE) = 50% for each trial.

In the case of amino acids with 20 possibilities, I used the approximation that all states are equiprobable, this is like a 20 sided die. For a polypeptide/protein of 1270 residues we “roll” each die and paint the side facing up WHITE with a number corresponding to what number die/roll it was. For an equiprobable distribution P(WHITE) = 5% for each trial. It’s still a binomial distribution. It won’t be if it’s not an equiprobable distribution.

In any case, if we use the equiprobable distribution as a first order approximation, the chances of another random polypeptide that polymerizes from a random soup being identical to another is 1 out of 20^N where N is the number of residues, which is astronomically remote.

The reason hexameric helicase can be formed for 6 identical polypeptides is that there is a genome that stores the recipe. But in a pre-biotic enviornment without genomes, this won’t happen.

@stcordova, the issue is that the first order approximation isn’t good. What is your second order approximation?


@stcordova, the issue is that the first order approximation isn’t good. What is your second order approximation?

One could take the most abundant amino acid, say the probability is 50% (which is outrageously high). Now let’s use 50% as the best case scenario an amino acid in one position will be the amino acid the same corresponding position in another polypetide.

The odds that it will duplicated exactly are 1 out of 2^N where N is the number of residues. So even under this generous assumption (it is wrong on the side of being generous), for a 1270 residue polypeptide being duplicated is still remote.

And this is only helicase, what of the other symmetrical protein complexes? One could assume of course, as is always done, that life arose in a totally different form. But that is pure faith and/or speculation, it isn’t empirical science. Empirical sciences says, “cells arise from other cells” in the present day. I’ve provided mechanical considerations why this is so.

Abiogenesis postulates different conditions that made life arise in the past, but never specifies them in a plausible frameworks. It’s just assertions that aren’t consistent with any direct observations, and often contrary to direct observations. That isn’t science.

That’s not true.

Since there are mechanistic hypotheses that make empirical predictions that have been tested, it is science.

1 Like

Sure. And what does this have to do with the price of tea in China?

1 Like

In this case, you are defining the outcome before the random process. This is the opposite of biology where you are defining the probability after the outcome.

If we flipped a coin 500 times and recorded if it was heads or tails, the order of H and T we get at the end is exceedingly unlikely, and yet it happened. If we calculate the probability of getting the specific outcome of H and T at the end, using your logic we would have to conclude that what we just did was impossible, yet it happened.

The overall problem for your arguments is that you are calculating the probabilities after they happen. This will almost always lead you down the wrong path.


He can look at an outcome and see if he can assign chance to that outcome. This is simply looking at data and assigning cause to the data you are observing.

Of course that can’t be done with any degree of fidelity unless you take into account the processes which produced the data. Something ID-Creationists NEVER do.

Can you calculate the probability of holding a royal straight flush 5-card poker hand without knowing if it was dealt directly or is the result of many rounds of discard and redraws?


@stcordova I’m not even worried about the approximation - I’ve set aside those other questions because I can honestly help make your argument a little better - or at least a little less wrong.

LLN is a statement about the probability of sums; Sums tend to be closer to some populationmean. Your example of 6 identical polypeptides, if we take 6 as a sum, might be a binomial or Poisson distribution, and we might then approximate those distributions with a Normal distribution (LLN). The statement of probability you are making is about a joint probability, and does not involve LLN.

1 Like

Then lets do that for the coins. Let’s toss a coin 500 times and record if it is a head or a tail. After 500 flips we can calculate the odds of getting that specific order of heads and tails as 1 in 2^500, or 1 in 3 x 10^150. Can you find any error in my math?

Yes, I can see it. The probability calculation is the chance of getting the first recorded sequence on the second 500 flips.

This is also the same as the odds of getting 500 heads as there is only one version of this in the population.

False. It is the probability of getting the first recorded sequence with no second set of flips, exactly what you are doing with molecular sequences.

1 Like

The probability of getting the first sequence is 1.

That’s also the probability of getting the molecular sequences we see, and for the same reason.

1 Like

Exactly. What is unknown is the mechanism that generated that sequence.

It is strange that you say this after being given the mechanisms multiple times now.