Is Functional Information Functional?

[Giltil (Gilbert Thill)]

@Rumraket

And then I really don’t see what practical benefit it has(this would be a criticism of Hazen and Szostak though). If one in (say) 10^12 sequences have the function of interest, or 1 in 10^30, why not just say that? Why bother with this FI stuff?

I would say for the same reason that information theorists have developed the concept of bit: transforming probabilities in bit of information presents several advantages.

[Rumraket (Mikkel R.)]

@Giltil

I would say for the same reason that information theorists have developed the concept of bit: transforming probabilities in bit of information presents several advantages.

You misunderstand me I think. I certainly see uses for the ability to quantify information in terms of bits*(there has to be some measure, a unit of information), I just don’t see what those advantages are in this case of estimating the rarity of certain functions in biology. What is the advantage of couching that in terms of bits of FI?

Measuring information content in bits is useful for determining things like storage capacity of a physical medium(hard drives, dvds, RAM, or even the storage capacity of a string of DNA), or transmission bandwidth if you want to find out how much information you can send in a limited amount of time(speed of your internet connection, WIFI, telecommunications and so on).

But if you want to tell me how much information the human genome contains, why not just tell me it’s total genome length? How much information is there in this particular protein? Just tell me how long it is.

If you want to tell me how rare such functions are in sequence space, just give me the actual number. 1.4\times10^{-12} of sequences (say). That way I don’t have to start working backwards from the FI score in bits, to get to the actual probability. It seems to just make it harder to get to the information that is useful or interesting. If you tell me it’s ~39 bits I now have to go solve an equation to get the probability. Just tell me it’s 1.4\times10^{-12}.

[Timothy_Horton]

@Giltil

I would say for the same reason that information theorists have developed the concept of bit: transforming probabilities in bit of information presents several advantages.

The problem is no one has near enough information to accurately compute a probability for any particular protein forming. Especially with the ridiculous assumption the IDCers use, that proteins all fell together from their constituent parts by chance.

Gpuccio and other IDCers are doing “science” bass ackwards. They started with the conclusion a Designer (their Christian God) created everything and are desperately looking for any numbers they can finagle to “support” their beliefs. FI is just one more tool they are misusing in pursuit of that goal.

[T_aquaticus]

@colewd

Good question. I still believe that gpuccio gave you the right scenario given a single AA change.

In that scenario FI becomes a subjective characteristic instead of an objective measurement as shown by your inability to measure FI in a non-functional sequence.

[T_aquaticus]

@colewd

The method does not use DNA as a measure of FI. Gpuccio does look at the change in DNA over time as a metric to see how much neutral mutation there is that does not change amino acid sequences.

The method could be easily applied to DNA. Perhaps you should read the 2007 paper by Hazen et al.

https://www.pnas.org/content/104/suppl_1/8574

In that paper they say FI can be measured in letter sequences. They also measure functional information in RNA polymers, so I don’t see why DNA polymers are any different.

@colewd

FI approximation is made by the constraint of the change to amino acids.

Yes, it measures the constraint AFTER the protein has evolved, not before.

[colewd (Bill Cole)]

@T_aquaticus

In that scenario FI becomes a subjective characteristic instead of an objective measurement as shown by your inability to measure FI in a non-functional sequence.

I think there is a solution here. Gpuccio got back to me and I am asking for additional clarification. After that I will get back to you and Rum.

[colewd (Bill Cole)]

@T_aquaticus

Yes, it measures the constraint AFTER the protein has evolved, not before.

Given a protein with a high level of constraint I remain skeptical “evolved” is a viable explanation for its origin.

[Faizal_Ali]

@colewd

I think there is a solution here. Gpuccio got back to me and I am asking for additional clarification. After that I will get back to you and Rum.

Why not just ask him to return to this group so he can answer for himself, rather than use you as an intermediary?

(OTOH, thanks for finally admitting there is a problem with your position that requires a “solution.”)

[Mercer (John Mercer)]

@colewd

Rum and Art are arguing for not being able to measure FI and Art and particular is arguing for a very low bit count of prp8 at 70 bits. They both realize 500 bits is a formable challenge.

No, there’s nothing to even suggest that. It’s a figment of your imagination. The point is that not only does FI fail as presented, the definition is worthless. Pointing out one in no way concedes the other.

@Rumraket already told you that you misrepresented his position. Whether you intended to do so or not, there’s only one ethical choice.

[Mercer (John Mercer)]

@colewd

Gpuccio got back to me and I am asking for additional clarification.

If it’s so clear, why would you need any clarification? Haven’t you been dismissive of those who dispute this, claiming that they don’t understand, both here and on UD?

How can you know whether others don’t understand if you so obviously don’t?

[Mercer (John Mercer)]

@Giltil

I would say for the same reason that information theorists have developed the concept of bit: transforming probabilities in bit of information presents several advantages.

I’m not seeing any. Perhaps you can demonstrate this alleged utility in the real world?

[Rumraket (Mikkel R.)]

Yeah same. I’d be happy to take correction here. I mean if there’s some practical use for the concept of functional information as defined by Hazen & Szostak, then fine, I will concede it’s utility if it can be shown. I just don’t see one as it is.

[Mercer (John Mercer)]

@colewd

Given a protein with a high level of constraint I remain skeptical “evolved” is a viable explanation for its origin.

But given how little you appear to understand of evolutionary biology, why should your skepticism carry any weight?

[colewd (Bill Cole)]

@Faizal_Ali

Why not just ask him to return to this group so he can answer for himself, rather than use you as an intermediary?

No, there’s nothing to even suggest that. It’s a figment of your imagination. The point is that not only does FI fail as presented, the definition is worthless. Pointing out one in no way concedes the other.

[Rumraket (Mikkel R.)]

@Mercer

No, there’s nothing to even suggest that. It’s a figment of your imagination. The point is that not only does FI fail as presented, the definition is worthless. Pointing out one in no way concedes the other.

Correct again. It is another misrepresentation by @colewd.

[Art (Arthur Hunt)]

@Mercer

Rum and Art are arguing for not being able to measure FI and Art and particular is arguing for a very low bit count of prp8 at 70 bits. They both realize 500 bits is a formable challenge.

No, there’s nothing to even suggest that. It’s a figment of your imagination. The point is that not only does FI fail as presented, the definition is worthless. Pointing out one in no way concedes the other.

@Rumraket already told you that you misrepresented his position. Whether you intended to do so or not, there’s only one ethical choice.

For what it’s worth, I did state that the FI for PRP8 might be around 70 bits. That comes from the fact that many different sorts of studies indicate that, as a general rule, this is a good ballpark for the FI of any functional protein, regardless of activity and/or length.

[colewd (Bill Cole)]

@Rumraket

Correct again. It is another misrepresentation by @colewd.

Rum you have been arguing for 4 years that there is lots of function in proteins.

[Mercer (John Mercer)]

@colewd

Rum you have been arguing for 4 years that there is lots of function in proteins.

That has no relevance to your blatant misrepresentation, Bill, but I am fairly sure that you know that.

You’ve claimed that Rum is a stand-up guy, and he says that you are misrepresenting his position. For an ethical person, the choice is obvious.

Simple questions from a biochemist:

1. How can biochemical functions be correlated with digital information?
2. How does FI account for partially-overlapping functions, a hallmark of biology that is predicted by evolutionary theory (and a major theme of 1990s mouse genetics), but is routinely ignored by evolution denialists?

[colewd (Bill Cole)]

@Mercer

That has no relevance to your blatant misrepresentation, Bill, but I am fairly sure that you know that.

You’ve claimed that Rum is a stand-up guy, and he says that you are misrepresenting his position. For an ethical person, the choice is obvious.

Simple questions from a biochemist:

1. How can biochemical functions be correlated with digital information?
2. How does FI account for partially-overlapping functions, a hallmark of biology that is predicted by evolutionary theory (and a major theme of 1990s mouse genetics), but is routinely ignored by evolution denialists?

John you have already dismissed the concept of FI being bogus. Whats left to discuss?

[Rumraket (Mikkel R.)]

@colewd

Rum you have been arguing for 4 years that there is lots of function in proteins.

Yes Bill, but what I am saying is a misrepresentation is your insinuation that Art and I are trying to say proteins have way less than 500 bits of FI because we think it is “a formidable challenge”.