The Argument Clinic

[Rumraket (Mikkel R.)]

@colewd

If the functional sequence space is smaller than the total sequence space what happens when you randomly change the sequence? Take your telephone and put a friends number down and randomly change it and see if anyone you know answers the phone to test this. How many changes on average would you fail to call one of your friends?

Imagine being so thick you write out that balderdash thinking it’s a useful analogy to evolution.

Already addressed your unimaginably stupid analogy two years ago. You never did come around to correcting it.

[Faizal_Ali]

@John_Harshman

Bill doesn’t lie. He just has very little control over what he types. He quite literally doesn’t know what he’s saying.

Yeah, and I shouldn’t need an avian phylogeneticist to remind me of this. Thanks.

[colewd (Bill Cole)]

@CrisprCAS9

That’s a claim, not a demonstration. And in any event, the size of the sequence space is irrelevant, only the function fraction.

They are both relevant. One is easy to calculate the other takes empirical data to try to make estimates.

The functional fraction needs to be enormous in all cases to make the evolutionary claims viable. The data from sequence comparisons does not support this idea.

[CrisprCAS9]

@colewd

They are both relevant

You can say they’re both relevant, but you’d just be wrong. Because you’re extremely, embarrassingly, incomprehensibly bad at math.

[colewd (Bill Cole)]

@CrisprCAS9

You can say they’re both relevant, but you’d just be wrong. Because you’re extremely, embarrassingly, incomprehensibly bad at math.

We are looking at a ratio. Ratios require both a numerator and a denominator. I would think a graduate student would know this. Where did you study math? What level did you achieve?

[Gisteron]

@colewd

If the functional sequence space is smaller than the total sequence space what happens when you randomly change the sequence? Take your telephone and put a friends number down and randomly change it and see if anyone you know answers the phone to test this. How many changes on average would you fail to call one of your friends?

I’m sorry, how many digits do phone numbers have where you come from? What’s the size of the phone number sequence space? Are some digit sequences technically incapable of being assigned a receiver? If no, are all digit sequences equally likely to be assigned a receiver? How large is the space of sequences actually available. How do we begin going about actually determining this?

[Gisteron]

@colewd

Ratios require both a numerator and a denominator.

Do they?

[CrisprCAS9]

@colewd

We are looking at a ratio. Ratios require both a numerator and a denominator.

You don’t know what a ‘fraction’ is, do you?

[John_Harshman]

@colewd

They are both relevant.

Your analogy is meaningless. You’re assuming that each mutation is a random sampling of sequence space, or equivalently that the functional sequences are randomly distributed in that space. Nobody thinks that either of those assumptions is correct. Mutations move to nearby points, and functional sequences are clustered.

[colewd (Bill Cole)]

@CrisprCAS9

You don’t know what a ‘fraction’ is, do you?

A ratio is a quotient between two numbers or two comparable quantities which is expressed as a fraction. There are two parts in ratio, one is the antecedent (numerator) and the other is the consequent (denominator)

Do you disagree with this?
.

[CrisprCAS9]

What matters is the fraction only. This is what you don’t understand. If the fraction is 0.1, it doesn’t matter if the sequence space is 10, 10 million, or 10^1000000. The fraction is the same.

So the sequence space does not matter. Only the fraction matters.

Do you understand now, or are you going to insist on continuing to be wrong?

[colewd (Bill Cole)]

@John_Harshman

Your analogy is meaningless. You’re assuming that each mutation is a random sampling of sequence space, or equivalently that the functional sequences are randomly distributed in that space. Nobody thinks that either of those assumptions is correct. Mutations move to nearby points, and functional sequences are clustered.

Protein families can exist in hugely different sequence space nowhere near each other.

https://www.uniprot.org/align/clustalo-R20231020-220128-0592-8363517-p1m/percent-identity-matrix

[Tim]

@CrisprCAS9

If the fraction is 0.1, it doesn’t matter if the sequence space is 10, 10 million, or 10^1000000. The fraction is the same.

Yes.

1/10 = 10/100 = 100/1000 = etc – they are all the same fraction – therefore you do not need to know the numerator or denominator to know the fraction.

(Agreeing with you, but just spelling it out in extra detail for Bill.)

[colewd (Bill Cole)]

@CrisprCAS9

What matters is the fraction only.

And how do you generate the fraction. Are these the silly argument going on in some unnamed graduate school?

[CrisprCAS9]

You might as well tell a cop he can’t arrest you for speeding because he doesn’t know how far you traveled.

Only the fraction matters. It’s not my fault you can’t accept this.

[Tim]

@colewd

And how do you generate the fraction.

Any way you like. For @CrisprCAS9’s and my fraction above, you could use 314159265359/3141592653590 if you chose, or any other denominator (with the correct numerator), or any other numerator (with the correct denominator). One or other is completely arbitrary.

[colewd (Bill Cole)]

@CrisprCAS9

Only the fraction matters. It’s not my fault you can’t accept this.

Sure but first you need to generate the fraction and you cannot make that calculation without a denominator.

For a 371AA protein such as alpha actin with a denominator of 20^371 possible arrangements how many functional arrangements (numerator) do you think are possible?

[Gisteron]

@colewd

@John_Harshman

Your analogy is meaningless. You’re assuming that each mutation is a random sampling of sequence space, or equivalently that the functional sequences are randomly distributed in that space. Nobody thinks that either of those assumptions is correct. Mutations move to nearby points, and functional sequences are clustered.

Protein families can exist in hugely different sequence space nowhere near each other.

So, in other words, you have no clue what John’s message means, but you are too itching to embarass yourself as to actually take a minute to learn anything about the topic.

[Tim]

@colewd

Sure but first you need to generate the fraction and you cannot make that calculation without a denominator.

False. A repeating decimal is a fraction – but lacks a (singular) denominator.

[Rumraket (Mikkel R.)]

@colewd

Protein families can exist in hugely different sequence space nowhere near each other.

UniProt

With members of each class being increasingly similar in sequence the further we go back in time. Idiot. That’s why they’re classified as belonging to the same superfamily in the first place. A 37% identity is a highly significant degree of sequence similarity for protein sequences almost 2000 amino acids long. They are not “nowhere near” each other.