Create a Protein with Your Mind

I wish I had more experience with biological systems, but I have been learning a lot of new stuff from the folks here on Peaceful Science.

My PhD work was on the photochemistry of chiral molecules and artificial molecular motors. I find the discussions around biological motors very interesting and for sure homochirality is also quite interesting. These days I spend the majority of my time as a science educator (at the undergraduate level) where I try to help students navigate the interactions of science and faith.

3 Likes

This thread is a good place to start.

A 4-pack of anything from here would be fine.

3 Likes

Holy forking cow. The irony

4 Likes

Yes. As accustomed as I am to creationist doublespeak, that one still pulled me up short.

1. Give or takeâŚ

70 bits?

Thatâs correct.

He is my logic. With 50% substitutability per position every position has a 50 percent chance of getting a problem insertion. For 10 digits this is the same as 10 binary digits or 2^10. So the formula is 2^ length of the sequence. What method gets you 70 bits?

This is not what BLAST is about. I suspect, though, that you are just using non-standard terms here that are unfamiliar to practitioners of the art, as it were.

I will say this yet again (for the 10^10^10^100000 th time) - the BLAST analysis @gpuccio is doing tremendously, completely under-reports the numbers of functional sequences. By hundreds of orders of magnitude. @colewd, until you understand this, you are just chasing rainbows. This goes for @gpuccio.

As has also been explained before, any estimate of FI that has a strict length dependence contradicts what direct experimental assessments of the frequencies of function in sequence space show. In other words, such an estimate is wrong.

I chose one end of the range (the end that favors the ID position) that more direct assessments of the fraction of functional sequences show (1 in 10^20) and did the appropriate log2 conversion. And rounded a bit so that we donât get too carried away - these are always very rough, back of the napkin calculations, nothing more.

5 Likes

So you are not challenging my assumption yet are claiming that out of around 5000 bits of total FI 4930 are functional. Itâs dinner out if you can support this in a reasonable wayđ
Functional means can splice well enough to reproduce.

How many bits do you think you need to get a successful 3D fold? We need combinations of hydrophobic and hydrophilic amino acids me thinks.

This is the same protein that is 99.9% preserved between humans and mice yet you claim that 99.9999âŚ % of the sequences can successfully splice.

Sorry @colewd, I have no idea what you just said. Almost nothing from your last post makes any sense. Maybe you can pick one topic and re-state things, so I can give another try.

Sorry.

2 Likes

Letâs table for now. Thanks for the discussion.

Okay. As I understand it, you are claiming that many sequences do exist that meet the minimal threshold function. And you are asking Bill to demonstrate that your claim is false. But this is unfair since proving a negative is notoriously difficult. This is precisely why a sane legal system does not require the defence to prove that a man suspected of murder didnât kill the victim, but require instead that the prosecution prove that he is guilty. So the burden of proof is on you, not on Bill.

You donât.

you are claiming that many sequences do exist that meet the minimal threshold function.

No, I am explaining that when you plug in such a number in the equation, YOU are claiming to know how many there are. This claim carries a burden of proof, and it isnât met by just sampling variants that exist in known life and extrapolating from that.

And you are asking Bill to demonstrate that your claim is false.

No, Iâm asking Bill to prove that HIS claim is true. The claim HE makes when HE says the number of sequences that meet the minimal threshold for function is the number he uses for that.

But this is unfair since proving a negative is notoriously difficult. This is precisely why a sane legal system does not require the defence to prove that a man suspected of murder didnât kill the victim, but require instead that the prosecution prove that he is guilty.

Yes and the case being assessed is that advanced by the prosecution, which seek to establish that sequence space is guilty of having exactly X number of sequences that meet the minimal threshold for function.

I am unconvinced that sequence space is guilty of that, so I want to see good evidence before I move to find sequence space guilty.

Without this evidence, I will have to conclude that the prosecution (Bill, Gpuccio, and whoever else plugs in numbers for the minimal threshold for function) have failed to meet their burden of proof.

So the burden of proof is on you, not on Bill.

No, the burden is on the person claiming to know how many there are(the person plugging in a number for M(E_X). I am questioning the claim by pointing out itâs advertised basis(there are these X number of sequence variants known in life) does not entail the stated conclusion(so we can extrapolate from that variation to estimate how many truly meet the minimal threshold for function).

4 Likes

The notion of âfunctionâ (in the mechanical sense) adds a confusion factor as it adds metaphysical notions of teleology along with math.

If something dies, a reaction doesnât go forward, so to that extent we might say something is âfunctionalâ but then the term is superflous.

A lot of objects in a typical household can be âfunctionalâ as a paper weight. For that matter a lot of random stones could do the trick. Minimal functionality isnât a good approach to trying to infer how difficult it is to make more complex function. Just because one can easily make a paper weight doesnât mean one can make a spark plug for a car via incremental steps.

The same issue applies to proteins. We can make probably trivially functional proteins (analogous to paper weights), but then there are multi-component proteins like this that arenât so trivial:

1 Like

Good point Sal. I think this is where we need more than just a âcomplexityâ spectrum to figure out whatâs going on here. One of the things that seems so convincing to me about common descent is that proteins (and the DNA sequences behind it) seem to be related in ways that seem unlikely if they were unrelated. In particular they seem related in a hierarchical way that indicates that there are varying amounts of relatedness.

2 Likes

Most certainly proteins of the same they are related in a nested-hierarchy. For example I built this one myself (an unrooted-neighbor joining tree) of the Cox1 protein:

And you can see the various creatures I chose. That said, the acknowledged difficulty is that the protein families themselves donât have a common universal ancestral sequence. This is a known problem.

ERRATA: That is a ROOTED NJ tree, not an UN-rooted tree, sorry

Can you give a reference that speaks about this supposed âacknowledged difficultyâ? What, exactly, is being acknowledged as difficult?

âŚ the protein families themselves donât have a common universal ancestral sequence. This is a known problem.

Is it? To whom, and why?

2 Likes

Well, we can start with Art Hunt who said in this forum that all proteins didnât descend from a common ancestral sequence, but that it isnât a difficulty. How can this NOT be difficulty?

This is akin and then saying âuniversal common descent is real, but ALL proteins donât have an ultimate common ancestor â the relationships are more like a orchard than a universal tree.â

For that matter, wouldnât you yourself agree:

ALL proteins donât have an ultimate common ancestor â the relationships are more like a orchard than a universal tree."