A Simple Gene Origination Calculation

Heh, this should create a little PANDA-moanium.

While I wish I could read the paper in it’s entirety, I must admit that the conclusion they came to is a little disturbing… though it depends on what your prior philosophical commitments are on the origin of life as we know it today.

They make it sound so simple. I wonder how many dead flies had to die before the genome was fully functional.

Jackanory.

Amazing coincidence, PaV. I was asking myself just a couple hours ago how many new functional genes come from duplicates of other genes and how many are de novo. (I think it was Michael Behe who criticized Kenneth Miller for overemphasizing the role of duplication. But if 44% percent of the new genes shared by Drosophila species arose by duplication, wouldn’t the percentage for a single species be higher?) I still wonder exactly how de novo formation works. And what is a non-functional gene? Does that mean it doesn’t yield a product, or does it mean that the product does not serve a function in the proteome? Does mutation sometimes turn non-functional genes into functional ones?

Is there a geneticist in the house?

F2XL, is your problem here that the abstract says indirectly that there are in fact non-coding regions in the genome that don’t serve a function? After all, turning a functional non-gene region into a gene would stop that region from serving its prior function. Are you concerned that this does not jibe with genetic entropy?

Didn’t Dr. Dembski argue in “Searching Large Spaces” and other writings that even one de novo functional gene is very unlikely to be discovered by random search? Or did he say that base-pair sequences he specified as a target were very unlikely to be hit? The abstract seems to address formation of any new genes that do something, not just genes that do what the researchers had in mind before looking at the genome. Isn’t this a much bigger target than Dr. Dembski considered?

Doesn’t this paper’s findings seems to support Darwinian theory?

PannenbergOmega:

Doesn’t this paper’s findings seems to support Darwinian theory?

The question is not ID vs. Darwin, the question is to what extent did the designer use the darwinian model — or as Behe so aptly put it, where is the edge of evolution?

There are two IDish things that glare out of this document.

1 - If this study establishes a base-line rate (which assumes no agency in the development of the fruitfly) then how did man get to be so genetically different from monkeys?

2 - The 11.9% De Novo genes, are these statistically reasonable, or are these scratch-marks from the designer’s tools; ie, evidence of agency?

This was covered in a commentary on the paper over at PT. (http://pandasthumb.org/archives/2008/06/more-on-the-ori.html#comments)

PaV is confusing various types of genetic change. There simply aren’t thousands of genes that humans have and chimpanzees lack.

Although PaV doesn’t site a specific study, I suspect he is referring to the number of genes which aren’t EXACTLY the same in humans and chips. Differences between two homologous gene sequence often arise from a single mutation event and usually ave no affect on an organism at all.

By contrast the drosophila study is measuring the rate of gene origin through a specific mechanism. Creating an entire gene takes significantly longer than making a small change to an existing gene.

By way of analogy PaV has found out the price of a gallon of water, filled up his car w/ 20 gallons of gas, and is now offering the clerk 2 dollars. This won’t work, because the price of a gallon of water isn’t the price of all other gallons of all liquids.

I’m surprised that such a simple and obvious error which has been corrected else where is of much interest to anyone.

The implications of this are awesome, and thus they will be entirely and predictably ignored by the Darwinian priesthood. It’s a very big problem not only for the putative hominid evolution line, but for primate and hominoid evolution across the board in all families. It’s worse than that of course (for Darwinians at least) -it’s a problem across the animal kingdom, and most obviously with mammalian evolution in all orders.

Goes to show just how overly generous Remine is being to the Darwinians in his appraisal of Haldane’s Dilemma. And even with the dice loaded in their favour, they are flummoxed.

Just one minor quibble with this blog entry (or perhaps major). Darwin’s theory of evolution does not claim that man evolved from the monkey. It claims that man and monkeys both evolved from a common ancestor which no longer exists. So the observation about how long it would take for man to evolve from the monkey is not relevant to the discussion.

Is a simple mathematical extrapolation of data obtained from an abstract really worthy of posting at Uncommon Descent?

This seems to be the intellectual equivalent of doing a movie review based on a trailer.

I wish people would stop using religious terminology (zephr) to insult Darwinists. Does that not reflect badly on religion?

Does calling atheism a “religion” really hurt atheism or does it hurt religion even more by associating the negative connotations of atheism with it?

The idea of a “Darwinian priesthood” is asinine. Who are you trying to insult here? Darwinists or priests?

I’m sorry but some of you people really need to think about what you’re saying and not just fly off the handle trying to insult atheists and Darwinists. It is more likely to hurt your cause than help it.

Off Topic: ID skeptic D’Souza rebuts Dawkins following Al Jazeera “debate”: http://townhall.com/columnists.....al-jazeera

PaV, thank you for sending me off to read about pseudogenes. As I understand it, a pseudogene is a DNA sequence that looks much like a known gene but does not function as a gene. Do you suppose a gene is called de novo if it isn’t a good match for any other known gene in the Drosophila species group? What I’m guessing is that all new genes not considered de novo arose through mutation of existing genes (changing function) or mutation of pseudogenes (introducing gene function). Does that make sense?

Biologists say that the rate of evolutionary innovation differs considerably from one species to the next. Would you please explain your rationale for extending the observed rate of appearance of new genes in a particular species subgroup to quite a different species?

I’m guessing that a small genetic change may yield a crucial first step into a new (maybe extended) niche. There may be many payoffs for genetic change in the new niche that are not available in the old niche. My un-expert hunch is that the rate of gene formation is related to the time a type has occupied its niche. This leads me to ask how long fruit flies have occupied their niche(s)?

Many people know that the rate of genetic change in humans is high. Relatively few know that it is also high in chimps (which are apes, not monkeys).

Again I ask, is there a geneticist in the house?

I used to have a reminder taped to the monitor of the computer I programmed: “Don’t guess!” It takes less time to look up the answer to something you’re not sure about or to write a little test program to see how things actually work than to guess and straighten things out later with debugging.

I feel kind of bad about guessing in my comments. I’m not lazy. The problem is that it takes some knowledge of a field to look up answers to questions in the field. I hope someone knowledgeable will point me in the right direction.

jdmack, we are not children here. We all actually understand the concept of the common ancestor.

matthew_ackerman, it is my understanding that the latest count of genes unique to humans, and without apparent source (de novo) is around 50. (I read that on this blog or at TT recently, but don’t have a citation.) 50 de novo genes in no more than 5 million years is a pretty good pace. Especially, consider that the lineage between man and CA probably had a generation rate of about 10 years, where the fruit fly’s generation rate is a few days.
1 - We’ve been evolving at quite a clip.
2 - Even the fruit fly’s rate of development of de novo genes seems awfully fast. I question whether its pace can be supported by statistical analysis.

The Zhou et al. abstract was discussed at the Panda’s Thumb on June 25. Some people there were concerned about the low rate of appearance of new genes. But according to biologist Ian Musgrave, who seems to have dug into Haldane’s Dilemma, there are no known de novo genes when chimps and humans are compared. See his comment for the explanation of what the differences are.

This also led me to Recent acceleration of human adaptive evolution, which appeared in the Proceedings of the National Academy of Sciences in December 2007. This is an extraordinary case where you not only get full-text for free, but can read an accessible explanation by the lead author, paleoanthropologist John Hawks.

Our evolution has recently accelerated by around 100-fold. And that’s exactly what we would expect from the enormous growth of our population.

His prefatory remarks make me feel not too dumb with my guessing and questioning above.

[A] very small fraction of the mutations in any given population will be advantageous. And the longer a population has existed, the more likely it will be close to its adaptive optimum — the point at which positively selected mutations don’t happen because there is no possible improvement. This is the most likely explanation for why very large species in nature don’t always evolve rapidly.

Instead, it is when a new environment is imposed that natural populations respond. And when the environment changes, larger populations have an intrinsic advantage, as Fisher showed, because they have a faster potential response by new mutations.

From that standpoint, the ecological changes documented in human history and the archaeological record create an exceptional situation. Humans faced new selective pressures during the last 40,000 years, related to disease, agricultural diets, sedentism, city life, greater lifespan, and many other ecological changes. This created a need for selection.

Larger population sizes allowed the rapid response to selection — more new adaptive mutations. Together, the the two patterns of historical change have placed humans far from an equilibrium. In that case, we expect that the pace of genetic change due to positive selection should recently have been radically higher than at other times in human evolution.

By the way, don’t discount the Hawks et al. article out of hand because of the fluffiness of what I cut and pasted from Hawks’ blog. They actually began by making predictions based on genetic theory, and then analyzed a large amount of genetic data. They tested their claim that human adaptive evolution has accelerated against a null hypothesis, and clearly rejected the null.

jdmack, (10)

You said,

Darwin’s theory of evolution does not claim that man evolved from the monkey. It claims that man and monkeys both evolved from a common ancestor which no longer exists. So the observation about how long it would take for man to evolve from the monkey is not relevant to the discussion.

It is true that standard evolutionary theory assumes that chimpanzees and humans (or monkeys and humans, for that matter) descended from a common ancestor which was not a modern chimpanzee. However, until the ancestor is found, it remains hypothetical, and the ancestor could be essentially a modern chimpanzee, or the ancestor, the modern chimpanzee, and the modern human could be genetically equidistant.

Even in the best-case scenario, where the ancestor is genetically precisely between modern chimpanzees and modern humans, the genetic distance between modern humans and the ancestor would still be half of that between modern humans and modern chimpanzees. Thus, the statement that “the observation about how long it would take for man to evolve from the monkey is not relevant to the discussion” is incorrect even if we replace “monkey” with “chimpanzee”. The observation does not eliminate the problem; it only decreases its magnitude by half at the most.

This objection reminds me of the claim that calculations cannot be made regarding the origin of life. Those making such a claim are simply trying to keep anyone from even attempting a calculation. They know that if such a calculation is attempted, their theory will appear ridiculous. Thus the spin that this area of science is beyond calculation. I guess that in these areas of science, only just-so stories are acceptable.

“They know that if such a calculation is attempted, their theory will appear ridiculous.”

Given appropriate assumptions, one can make any theory look ridiculous.

Mr King:

First, kindly inform us as to how calculations such as those summarised here are premised on dubious assumptions.

Next, let us look at PaV’s deductions above, as a back- of- the- envelope style, order of magnitude estimate:

As finding number 6 in the abstract, the authors write: “the rate of the origin of new functional genes is estimated to be 5 to 11 genes per million years in the D. melanogaster subgroup.”

Noting that Drosophila melanogaster has 14,000 genes (a very low gene number), the simply calculation is this: 14,000 genes/8 [NB: 8 is mid-point to [5, 11]] new functional genes per million years= 1.75 billiion years for the formation of the fly genome. This, of course, assumes that somehow the fly is ‘alive, and reproducing’ the entire 1.75 billion years—-this, without the aid of a full-blown genome. If we apply this to the monkey/human difference which, IIRC, is about a 1000 genes, then using this same rate, it would take 200 million years for man to have evolved from the monkey. This published rate for new functional gene generation cannot be good news for Darwinists.

So, kindly explain: where are the dubious assumptions in the above, and why are they dubious?

GEM of TKI

PaV, I believe that de novo genes are genes that are not modifications of pre-existing genes. For instance, a duplication could make an additional gene, but it would not be de novo. There are a variety of patterns — insertions, reversals, etc., which make new genes, but genes that are not de novo. If some non-coding “junk” DNA suddenly got incorporated as a working gene (not infeasible, it just needs a mutation to make a start marker, and it needs to make some functional sense) then we would have a de novo gene.

However, this article says that in fruit flies about 11% of the new genes are of the de novo variety. If humans have 689 new genes, the fruit fly experience would imply that we have about 75 de novo genes. Other sources suggest that we have between 50 and 100 de novo genes. The report that we have NO de novo genes would be quite surprising in light of other reports and in light of the fruit fly study.

Should we thus conclude that whereas species of fruit flies are separated by whole genes either being present or not, this isn’t the case when it comes to apes and men. Does this sound reasonable at all?

Yes. There are many more Drosophila than humans at present, and the human population has exploded only recently. It seems safe to say that the number of Drosophila has been huge throughout the past 6 million years. Furthermore, if you accept the generation rates someone gave above of 3 days for Drosophila and 10 years for humans, Drosophila have gone through about 1200 more generations over the past 6 million years than humans have. The Hawks et al. paper I linked to above would lead us to expect many fewer beneficial genetic changes in humans than in Drosophila, and it is “reasonable at all” to extend this to new genes.

Actually, I’m not taking into account the environmental shift of humans, and the presumably stable environment of fruit flies. This doesn’t change my statement that it would be reasonable, given what we know, for there to be no de novo genes when humans and chimps are compared. It simply means that you definitely cannot take rates and percentages from one species group and use them to make estimates for a species under quite different circumstances.

CEC09, (25)

You did mean 1200 times more generations for fruit flies than for humans, I hope.

Daniel King, (21)

You said,

Given appropriate assumptions, one can make any theory look ridiculous.

That depends on how you define “appropriate”. If you mean assumptions chosen to embarrass a theory, that is debatable, but just barely possible for the theory of gravity and its more sophisticated relative, the general theory of relativity. Origin of life theory, however, can be embarrassed even by using rather straightforward, and even generous assumptions, as I noted here. (Comment 49. Note: read carefully before replying, to avoid personal embarrassment.)

CEC09:

I read the abstract, and I really think that it is impossible to comment on it without reading the whole paper, which I can’t.

I will therefore make a few remarks about the above discussion, and especially about some of the many points you raised.

You say (#4):

“I still wonder exactly how de novo formation works”.

Nobody knows how it works. We just know that we find in species genes which have no significant homology with other known genes, even in a very general way. Those are named “de novo genes”. Nobody knows how they arise. They are certainly vastly beyond any probabilistic resource. In essence, any kind of gene must have been, in the beginning, a “de novo” gene. Then, gene families are observed, where homologies can be found between the components of the same family. But nobody knows how “de novo” genes first arise, and in reality nobody knows how and why genes which have some homology are related. Even accepting some form of descent from one gene to the other, in most cases the mechanism of change is not known. The proposed mechanism of random variation and some form of selection is comnpletely impotent to generate those kinds of effects, as demonstrated by all the various calculations we have often discussed here at UD.

You say:

“And what is a non-functional gene? Does that mean it doesn’t yield a product, or does it mean that the product does not serve a function in the proteome?”

Usually we call “gene”, or more precisely “protein coding gene”, those DNA sequences which have the characteristics of being transcribed and translated. That can be an inference from the structure of the DNA sequence, or we can have the evidence of the transcription and translation (in other words, we can know the mRNA or the protein). But the number of genes is often merely deducted by the analysis of the DNA sequence, and many studies about homologies are conducted in the same way. Obviously, when the protein is known, and so its function, the gene can be studied in a more realistic way.
Pseudogenes are genes which are very similar to functioning genes, but are no more functional. I think that usually they are neither transcripted nor translated, but I would not take anything for granted in this very difficult subject: anything we actually know about non coding DNA (including pseudogenes) is in my opinion highly provisional.
Finally, I am not sure, but I don’t think we have evidence of translated proteins which have no function in the proteome, and which are there, waiting to be magically transformed by random mutations to acquire a purpose. Usually, a cell is a very crowded place, and each protein has its role.

You say:

“F2XL, is your problem here that the abstract says indirectly that there are in fact non-coding regions in the genome that don’t serve a function? After all, turning a functional non-gene region into a gene would stop that region from serving its prior function.”

I don’t think the abstract says that. There was a thred some time ago about a much more specific work about the formation of a de novo gene (I don’t remember the name of the thread or of the article), and the possibility was discussed that a de novo gene originared by mutations from a segment of non coding DNA. The interesting thing is that the supposed “progenitor” sequence seemed to be transcribed and functional. So, your question is legitimate, and indeed it is the same question which i made in that thread: if the original non coding DNA was transcribed and functional, what happens when it gives rise to a protein coding gene? (In the case discussed, there was no evidence of duplication of the original sequence).
Moreover, the results of the ENCODE project seem to show that practically all the genome is transcribed. That does not prove that it is functional, but it is a good step in that direction.

You say:

“Didn’t Dr. Dembski argue in “Searching Large Spaces” and other writings that even one de novo functional gene is very unlikely to be discovered by random search?”

Yes, he did, and it is. If we really find de novo genes in species, and we do find them, and will find them ever more, that is only a good reason to ask ourselves how those genes ever did arise. The only reasonable answer is design (see next point).

You say:

“Or did he say that base-pair sequences he specified as a target were very unlikely to be hit? The abstract seems to address formation of any new genes that do something, not just genes that do what the researchers had in mind before looking at the genome. Isn’t this a much bigger target than Dr. Dembski considered?”

That’s a very important and pertinent question, and often a cause of misunderstanding. The subject is vast, but I will just outline some aspects:

a) Nobody really knows how big the target of functional proteins is, but we have all the reasons to think that it is very, very small compared with the immense search space of all possible protein sequences. Let’s make an example, amd take the generic search space of all proteins of 100 aminoacids (which are indeed very small proteins). The whole search space is about 10^130. That’s something!
But how many of these proteins can be functional? Nobody really knows, but even if we take “functional” in a very big sense, that is “able to have some useful function in some kind of living being”, still we have to remember that there are a lot of constraints, some known, some unknown, to that result: proteins have to be able to fold, and to fold in some ordered way, usually corresponding to some fundamental 3D structure. The folded protein must have domains, and active sites, which can exert some function, usually interacting with other proteins or with other organic molecules. In other words, there are all kinds of arguments, some of them also experimental (see for instance the interesting field of protein enineering), to believe that only a minor subset of proteins can be functional.

But how minor? We have to realize that, with a search space of 10^130, even a very big number is really minor. Let’s pretend, just to discuss, that 10^30 proteins of 100 aminoacids are in some way functional: they can fold in some orderly way, and they can have some biologic function, as enzymes or in other ways. 10^30 molecules would seem a very big target, and it really is. But, in a search space of 10^130 sequences, there would still be a probability of only 10^30/10^130, that is of 1:10^100, of finding that kind of target by chance in a random search. And that is still a probability vastly beyond any reasonable biologic resource (yes, I know, that’s not beyond Dembski’UPB, but we have to be realistic: that kind of UPB is only an extreme reference, useful in a theorical discussion, but if we discuss real life and biological systems, probably any level around 10^30 or 10^40 is more than enough).

b) But the above reasoning about “any protein which could have some function” is still only an extreme, theorical argument. In reality, given a biological context, only a very minor subset of generically functional proteins can have an “useful” function. In other words, the more a context is complex, the more a new function becomes difficult to find by chance. Let’s remember that most proteins, in a cell, are functional only because they interact precisely in complex, and I would say irreducibly complex, networks of other proteins, and many of those networks are really abstract in nature: they transmit information from cell to cell or inside a cell, they regulate other functions in very finely tuned ways, and so on. Inserting a new element in such complex and finely tuned contexts really requires a very precise choice of functional proteins.
So, the problem is not that a cell should give rise to any possible functional protein (which, as shown at point a) would anyway be impossible), or to the protein Dr. Dembski or anybody else defines. The cell has to find some specific functional protein which can be useful in the context of the existing biological network. A protein which can be useful in a species will not be useful in another one, and even minor differences can create enormous losses of functionality.
So, given that, would you still believe that a specific organism, in a specific context, can really choose in a target of 10^30 functional proteins of 100 aminoacids (which was in itself a very generous exaggeration)? Or should we reduce the target to 10^10, or to 10^6?

The real fact is that it doesn’t matter. A search of 1:10^100, for all practical purposes, is as impossible as a search of 10^130, if we don’t incorporate information about the desired result in the search. And the only principle which can incorporate that kind of information is design.

CEC09:

you definitely cannot take rates and percentages from one species group and use them to make estimates for a species under quite different circumstances.

I fail to follow your reasoning here. While I would agree that you cannot use the rate of evolution of one species to determine the rate of evolution of another species down to multiple decimal places. However, we are discussing orders of magnitude here. We should be able to get some sense of orders of magnitude accross different genetic lines. This idea that we can know nothing from evidence extrapolation is just as silly as the idea that we can determine with precision via extrapolation. We need to get beyond this simple black and white thinking.

bFast, extrapolation is regarded as highly dangerous throughout science, not to mention statistics. There are many examples of the errors it leads to.

One of the reasons I introduced the Hawks et al. paper to the discussion is that it gives very strong evidence that the rate of human evolution has accelerated by TWO orders of magnitude in recent times as a consequence of the population explosion and shifts in living conditions. In other words, extrapolation of the human present to the human past, or vice versa, leads to grossly incorrect conclusions.

When within-species extrapolation of the rate of genetic innovation is so strongly contraindicated for humans, and has been shown to be strongly related to population size, how can you possibly defend extrapolating results for a species group with populations that have long been much larger than the present human population to the human species?

My objection to extrapolation of the numbers in the Zhou et al. abstract is not merely a matter of principle. Rigorous scientific study of the human genome clearly tells us not to extrapolate. This is not just one of my unexpert guesses.

You did mean 1200 times more generations for fruit flies than for humans, I hope.

Thanks, Paul Giem. I was a bit rushed. Ten years is about 1200 times as long as 3 days.

Even in the best-case scenario, where the ancestor is genetically precisely between modern chimpanzees and modern humans, the genetic distance between modern humans and the ancestor would still be half of that between modern humans and modern chimpanzees.

I’m glad you brought this up. It’s more reasonable, though somewhat simplistic, to think of chimps and humans as separated by 12 million years than by 6 million years.

I’d like to see the whole paper.

Maybe someone who knows more about fruit fly research can tell me if any “new functional genes” have actually been observed to arise in all the generations of Drosophila studied?

If even one has been observed, it seems that a rate could be calculated, rather than having to be estimated.

I none has been observed, how do we know it ever happens?

Paul Giem #27:

“(Comment 49. Note: read carefully before replying, to avoid personal embarrassment.)”

Thanks for the warning. That is such a kindness. You are a Christian in the best sense.

CEC09:

the rate of human evolution has accelerated by TWO orders of magnitude in recent times as a consequence of the population explosion and shifts in living conditions.

1, how recently has there been a significant shift in living conditions for human/humanoid species? 10,000 years? Even a 100 times increase in evolution means little with this much time accounting for it. The vast lions share of the 6 million years of interest is at the slower pace. What of the fruit fly, has its environment not been affected by modern farming practices. I’m sure they love fruit orchards.

We also see that population (which until about 10,000 years ago was miniscule for the human lineage) benefits evolution. Fruit flies vastly outnumber us, don’t they? They certainly vastly outnumbered our ancestors 1/2 million years ago. Every which way you turn, evolution should be happening vastly more slowly in humans than in fruit flies.

My calculations suggest that, per generation, humans have evolved over 4 orders of magnitude faster than fruit flies. That, in light of the fact that there are many more fruit flies than humans (especially averaged out over the 6 million years.)

# generations in 6 million years:
Human 600,000. Fruit fly 600,000,000.

# new genes during this time:
Fruit fly: 50 (8 per mil * 6)
Human: 689

# new genes in human during this time if they evolve at fruit fly’s rate (per generation) 0.05

Difference in pace: about 14,000 to 1.

dacook, “I none has been observed, how do we know it ever happens?”

If one species of fruit fly has a gene that all other species don’t have, a new gene must have come from somewhere. It does not require observing the new gene appearing to determine this. However, this observation does not give us the cause — was the new gene the product of dumb luck or of an active agent.

The real fact is that it doesn’t matter. A search of 1:10^100, for all practical purposes, is as impossible as a search of 10^130, if we don’t incorporate information about the desired result in the search. And the only principle which can incorporate that kind of information is design.

If your assumptions were right it would be unlikely that your very own immune system would ever generate any functional antibody.
BTW, you may read a little bit on immunoglobulin gene diversification in birds that heavily relies on pseudogenes and gene conversion. You will find out that birds and those mammals that employ the same mechanism to generate antibody diversity can indeed produce antibodies against antigens that they or their ancestors never encountered before.

gpuccio:

Those are named “de novo genes”. Nobody knows how they arise. They are certainly vastly beyond any probabilistic resource.

How can you establish a very low upper bound on the probability of an event when nobody knows its cause(s)?

Moderators:

Re Mr King at 33 above.

He has no permission to use my personal name in this blog or elsewhere. This is relevant first as there is a known asymmetry between ID supporters and opponents, given the issues raised in e.g. Expelled. Mr King either knows, or should know that.

Secondly, in my case, I have found that use of my personal name and/or contact opens up to spam attacks. (I have retained in my personal site, ways to find my name and contact, but that is for responsible behaviour. What was done above is at minimum grossly irresponsible, and highly disrespectful.)

I must therefore ask Mr King, on pain of further, more serious appeal to the Moderators, to cease and desist.

GEM of TKI

sparc:

First of all, I really believe that my assumptions are right, and I invite you to tell where they should be wrong. In a sense, they are not assumptions at all, but very simple facts about probability.

Regarding the immune system, the scenario is completely different. Primary antibody diversification is a process which uses random variation very intelligently targeted to generate a repertoire of basic antibody specificities to cover, at a low specificity level, a search space which is very big, but not immense, referring to possible epitopes in nature (an epitope is a very small aminoacid sequence, usually a few aminoacids, or up to ten -fifteen). Even so, the basic repertoire is very unspecific, and can ensure only a low level interaction with possible epitopes. Antibody maturation “after” primary response, instead, is a typical process which utilizes random variation very intelligently targeted plus very intelligent selection to increase the specificity of the immune response. Indeed, the process utilized here is the same as used in modern protein engineering: the results of targeted random variation are “measured” against the original epitope, and intelligent selection takes place (obviously, here selection includes very specific informatioon about the target, that is the epitope itself, and is therefore very efficient).

So, as you can see, there is nothing in what we know about antibody generation which is inconsistent with my “assumptions”. Antibody generation is a perfect example of intelligent engineering using the realistic resources of probability. It is therefore perfectly natural and reasonable that the immune system of birds or mammals can “produce antibodies against antigens that they or their ancestors never encountered before”.

CEC09:

“How can you establish a very low upper bound on the probability of an event when nobody knows its cause(s)?”

It’s very simple. The probability of an event is calculated for the hypothesis that a random process of variation is the cause. That is the assumption in darwinian evolution, or whatever you want to call it.

If the cause is not random variation, but some other causative process (like laws of necessity or design), no calculation of probabilities makes sense. Probabilities are calculated for random events. For events which have a specific cause (necessity) the probability is always 1 (the event has to happen). The point is, as we have often discussed here at UD, the kind of information we observe in biological systems (CSI, or FSCI) simply “cannot” be the result of laws of necessity. So, if you can exclude random processes as a cause (and that’s why we calculate their probabilities), then design is the only known causal process which gan give that kind of result.

So, to sum up: we don’t know how a new gene originates: we are not sure about the causal mechanism and we certainly don’t know the details of the process. But, about the mechanism, we can make some sound inferences:

a) It cannot be the result of necessity (as often discussed, necessity cannot give that kind of complex information).

b)It cannot be the result of random variation (probabilities are absolutely against that).

c) It can, definitely, be the result of intelligent design (we see that happening every day in human artifacts).

That’s why, as we always say, design is at present and by far the best explanatory theory for CSI in biological beings.

gpuccio:

From the perspective of most geneticists and evolutionary biologists, “evolutionary search” is an oxymoron. The term search immediately suggests a goal, contradicting their belief that biological evolution is non-teleological. As Norbert Wiener pointed out in Cybernetics, evolutionary adaptation is much like learning in individuals and in cultures. But it’s easier if we stick with terms familiar to you, provided no one uses them to beg the question of teleology.

10^30 molecules would seem a very big target, and it really is. But, in a search space of 10^130 sequences, there would still be a probability of only 10^30/10^130, that is of 1:10^100, of finding that kind of target by chance in a random search.

The 100-amino-acid protein is a very old example, predating Dembski. No one claims that there has been a search of 100-amino-acid proteins. No one even claims that there has been a search of length-306 strings of nucleotides (I’m including START and STOP codons). Evolutionary search is on the space of genomes (and even that is a modeling fiction, ignoring epigenetic information). Furthermore, “random search” is highly misleading in suggesting “uniform sampling with replacement.” In fact, the nucleotide mutation rate is quite low. Genomes tend to grow, not shrink, in size due to such phenomena as gene duplication and retrotransposition. And these introduce new genetic regions that are anything but initialized uniformly at random. They start out with much or all of the information of functioning genes. Other interesting biases in the search arise from the fact that some amino acids are coded for by as many as six codons, and some are coded for by just one. AUG codes for both START and methionine, with interpretation according to context.

These are just a few aspects of genetic change tossed off by a computer geek too tired to see straight. But they’re enough to make analysis incredibly difficult for a Rip Van Einstein on his second cup of coffee.

A key problem, in my mind, is that we know very little about the topology of the set of functional genes. Do functional genes tend to cluster, in some sense, in the space of nucleotide sequences? What is the probability that a short, functional gene homologous to none other will grow into a slightly longer functional gene? This could happen if an AUG codon arrived, by mutation or some other cause, just ahead of the existing start-AUG, changing the original start-AUG into methionine-AUG. What I’m driving at is that you cannot fix the length of codon sequences in your analysis. If I recall correctly, there are 10-amino-acid artificial proteins.

Another important consideration is that the search over genomes may produce new genes in many overlapping regions. To get the flavor of what I’m talking about, compare these two questions: What is the probability that four tosses of a fair coin will yield the outcome HTHH? Now what is the probability that sixteen tosses will yield an outcome with HTHH as a subsequence? The regions of the genome that might give rise to genes are much longer than genes, so the analysis should look more like an answer to the latter question than the former (though, again, the sequence length should not be fixed, and sampling is not uniform). In other words, if you restrict your focus to a certain subsequence of nucleotides, you greatly underestimate the probability that a functional gene of a certain length will arise.

Now, on matters of substance:

1] “Assumptions”

Mr King, we are dealing with a known context, not an abstract discussion of logico-mathematical method.

In that context, you have invited by pretty direct inference, the conclusion that the sort of calculation above by PaV and/or the design thinkers more broadly, is unwarranted. Above, I therefore challenged you to substantiate — which is plainly not an act of pride but of asking for the reasons for YOUR hope.

In response, sadly,you have resorted to disrespectful behaviour as I just had to note, and have thereafter tried to improperly reverse the burden of substantiation.

Worse, in 22 above, I actually provided a link to my always linked substantiation in outline of the sort of configuration space-based calculation that is often presented in an ID related context. (So much for the remark: “As an exercise in humility, you might answer your own questions above…” For, the “exercise” has been long since done, and you as a long time participant here know or should know that.)

That calculation, anchored in the inherent nature of digital systems, raises serious challenges regarding the reasonableness of the idea that chance + necessity within the gamut of the observed cosmos, could give rise to the sort of discrete state information entities that lie at the heart of cell based life.

You have provided neither a serious response on the merits nor a link to such. [Onlookers: PG's link to TO is illustrative of just how unserious the response from the evolutionary materialist side too often is. That's why he linked it without further comment.]

The logical conclusion is that you are diverting from the substantial matter, and are taking up personalities instead. That speaks volumes on your real estimation of the substantial weight of your side’s case on the merits.

2] The TO assertions:

TO of course, first, does not wish to acknowledge that anyone other than “creationists” raises serious questions; never mind teh presence of men like Fred Hoyle — a distinguished scientist and a life-long agnostic — in the immediate context:

Problems with the creationists’ “it’s so improbable” calculations

1) They calculate the probability of the formation of a “modern” protein, or even a complete bacterium with all “modern” proteins, by random events. This is not the abiogenesis theory at all.

2) They assume that there is a fixed number of proteins, with fixed sequences for each protein, that are required for life.

3) They calculate the probability of sequential trials, rather than simultaneous trials.

4) They misunderstand what is meant by a probability calculation.

5) They seriously underestimate the number of functional enzymes/ribozymes present in a group of random sequences.

Let’s look a bit closer:

a –> Events are caused, per massive observation immemorial from the day of Plato, by chance and/or mechanical necessity and/or intelligent action. If we, ex hypothesi, rule out intelligence, that leaves chance [for high contingency situations or aspects -- which side of the die is upppermost] and necessity [for low contingency situations or aspects -- if dropped the die will fall].

b –> Cell based life, notoriously, rests on macromolecules assembeled from monomers, and cells are thus discrete-state systems, with very high contigency states. For the component DNA and RNA, 4^N, where N for OBSERVED — note TO’s artful, dismissive use of “modern” to dodge this key point — life systems starts at 300 - 500,000. For proteins, 20^N, with 100 - 500 being typical chain lengths. The relevant config spaces therefore START at ~ 10^130, and go up from there, WAAY up from there.

c –> Biofunctional states of DNA and proteins are highly constrained, the former by a code process and the onward link to proteins, which as GP outlines abiove, are massively functionally constrained. Sensitivity of funcitonality to perturbation of chain structure simply underscores this point.

d –> Thus, we easily see why the observed functional systems are highly vulnerable to perturbation. (AKA: Why is there no market for deliberate exposure to mutation inducing irradiation? Why is it that most cases of observed microevoltuion are by loss of function?) That immediately means that the probability of gatting to such entities in biofunctional clusters per one variety or another of Darwin’s soup in a prebiotic pond or geothermal vent or a comet, etc, is vanishingly small on ther scale of the observed cosmos. There are many upper bound estimates that show that, very strongly; these are the probability calcs that TO wishes to dismiss.

e –> So, contrary to assertion no 2, as GP shows above, and taking in the calculations as a class, there is not any imposition of a dubiously arbitrary bound on proteins etc. Instead, there are reasonable estimates of models, with a consistent pattern of results, one that cuts across what TO wishes to sustain. [Much, much more reasonable than say the basis for the GCMs that have so much of the world in a tizzy over climate change.]

f –> 3 is either a bald lie or reveals grossest incompetence. For, the Dembski type bound is based on the number of quantum states attainable by the observed cosmos as a whole, some 10^80 partticles, sifting through combined states at the rate of one combination of states for 10^80 or so particles, every 10^-43 seconds or so. That is as “simultaneous” AND as “sequential” as it gets. Worse, the very FIRST ID technical level book, TMLO, used an equilibrium state thermodynamic discussion, which is plainly and explicitly based on an overly generous concentration of monomers in a planet-scale prebiotic soup. And the issue is the monomer to biofunctional polymer transition, not hte assembly of a bacterium de novo out of the monomers.

g –> Furtehr to this, probability calculations, of course are usually based on observed frequencies, or on the Laplacian equal chance of alternatives hypothesis, perhaps biased by some non-uniformity in probability distributions. The equiprobability of microstates is foundaitonal to statistical thermodynamics, a highly successful sceintific discipline. As my point 6 app 1 the always linked therefore shows, this sort of approach is a more or less standard way to look at the relevant probabilities. (In TMLO, which specifically addresses polymerisation [so the simple chemicals to bacteria assertion in TO is a strawman misrepresentation!], polymer expert Bradley actually addresses the Kenyon Biochemical Predestination chemical-bias argument through a statistical study and so convincingly was it overturned that Kenyon took advantage of writing the foreword to publicly recant.) In short, TO’s 4th assertion is little more than a slander.

h –> Finally, one may assert away all s/he wants that there is a large number of combinations of monomers that are “functional.” The vastly yet larger number that are not, immediately reduces the suggestion to absurdity — the smallest known functional genomes are of order 300 - 500 k.

i –> That sets up a config space of minimal order 4^300,000 ~ 9.94 *10^180,617. With a generously assumed 10^1,000 islands of biofunctionality, each with 10^1,500 functional configs to choose from, we would be looking at 1 in 10^178,000. Negligibly different from zero probability on the gamut of our observed cosmos, or even for an array of 10^500 such sub-cosmi in some sort of brane sheet or the like.

j –> “Generous”? Indeed: 10^1,000 or 1,500 is vastly larger than the number of not only bacteria — to a first approximation, every lifeform is a bactrerium! — that could have ever lived in our cosmos, but of the proteins or DNA or RNA strands that could have ever formed. For, the observed cosmos could enfold up to only 10^150 quantum states across its lifespan.

k –> So, it is safe to conclude that a random walk through such a config space starting at an arbitrary initial point — what a prebiotic soup would do — is maximally unlikely to ever find the shores of any single island of funcitonality, much less get to apply any hill-climbing algorithms such as some vcariety of prebiotic natural selection. [Which itself is a questionable notion.]

l –> And, if there is an assumed natural law that programs the emergence of life, we should note that the term “natural law” relates to natural regularities, i.e. situations that are of low contingency: if you have heat, fuel and air, you have a fire. But, life systems are just the opposite of low contingency. That would leave only chance or agency as the credible explanations. [Cf the config space taken up by this post's character sequences -- chance or agent or law?]

m –> Worse yet, the future discovery of such a “life-program law” in the face of such a vast config space and resulting set of oterwise accessible contingencies, would imply pretty directly that the cosmos was programmed at its origin to create life. Programs of course require programmers.

In short, there is a serious case on the merits to address. And, neither DK nor TO have taken it seriously.

GEM of TKI

PaV:

Thanks

GEM of TKI

PS: I would love to see a case of an observed life-form using small sets of short-length genes, RNA and proteins that in aggregate does not get us into trouble relative to the UPB as I just discussed [and pardon the typos -- ouch!].

DK, Sparc and CEC09, might you be able to provide us a link?

Failing such, could you provide us with cogent — non question begging — reason to infer that chance + necessity with no intelligence is a superior explanation to agency for the origin of the functionally specified, complex information in observed life forms?

Or, just, can you give us an example of FSCI beyond the Dembski type bound i.e. ~ 500 - 1,000 bits of information storage capacity, that — per direct knowledge of the causal process — originated by chance and necessity only? [In short, can you show us that the explanatory filter is an empirical failure, instead of the reliable test for intelligence that we have claimed, per massive observation on the origin of cases of FSCI? (Just as, an observed perpetual motion machine would at once disestablish the laws of thermodynamics.)]

CEC09 (#40):

You say a lot of correct things which do not change a comma of what I have said. The 100 aminoacid protein is obviously a standard example, just to try a computation. I have to recall that most proteins are much longer, and much more complex.

It is obvious that the length is not fixed: that makes things even more difficult, and the search space even bigger.

It is possible that sometimes we can pass from one functional gene to a similar functional gene, with some slight change in function: that’s exactly the rare situation which could in theory occur by random mutation, and the few artificial examples of darwinists (see corticosteroid receptor experiments) are of that kind.

But are you sayin that all functional genes are similar? Or that all types of proteins (and genes) cluster in the search space? That’s simply absurd. Proteins are extremely different, they fold in completely different ways, they have very different domains and functions. Some proteins have completely no homologues. Even the best known protein families (see myoglobin) exhibit striking variations, often of a compensative nature, so that very different sequences may fold in a similar way.

So, it is not true, imo, that “we know very little about the topology of the set of functional genes”. We certainly know little, but we definitely know something, and nothing of what we know encourages the strange and fantastic topologies which darwinists like to imagine, like the smooth transitions from one cluster to another in the immense ocean of possibilities.

It is certainly true that it is difficult to calculate the probabilities exactly, but that does not mean, in any way, that we cannot have a definite idea of their order of magnitude. Frankly, objections like the one you make about the asymmetric nature of the genetic code seem really irrelevant. I made my calculation from aminoacid sequences, and that was just a generous simplification. If you look at DNA sequences, you must add, to all the possibilities of protein sequences, also all the sequences which would never be transcripted and/or translated, and I can’t see how that would help your argument.

Finally, I see in your reasoning a common misunderstanding: when I use the term “random variation” I refer to any possible variation, and not only to single nucleotide substitutions. You can put in any mechanism which is non directed, be it deletion, duplication, inversion, insertion, and so on. The result does not change. Any of this act is random, and the result is a random sample of the whole search space.

So, it’s not true, as you say, that ““random search” is highly misleading in suggesting “uniform sampling with replacement.” It does not suggest anything like that. It only suggests that the cause of variation is not directed, and has no informational relationship with the function which is the final result of the variation (is that non teleological enough as a substitute for “search”?).

It is not even necessary that all results are equally probable. Unless you can demonstrate that there is some law of necessity which determines special probability distributions which favor function, any real probability distribution, having no informational relationship with function, is blind to the result’s functionality. So, the search is still random. In other words, if variation is caused by laws which have no relationship of any kind with a special informational pattern which allows function, the emergence of that pattern is possible only either by chance or by design. If chance is too unlikely, as it definitely is in the case of biological information, then only design is a reasonable answer.

GP

Well said again.

GEM of TKI

A thought:

Here is Wiki on Drosophila, as a little 101-level context will I think be helpful:

Drosophila is a genus of small flies, belonging to the family Drosophilidae, whose members are often called “fruit flies” or more appropriately vinegar flies, wine flies, pomace flies, grape flies, and picked fruit-flies, a reference to the characteristic of many species to linger around overripe or rotting fruit . . . D. melanogaster, has been heavily used in research in genetics and is a common model organism in developmental biology . . . The entire genus, however, contains about 1,500 species and is very diverse in appearance, behavior, and breeding habitat . . . .

Drosophila are found all around the world, with more species in the tropical regions. They can be found in deserts, tropical rainforest, cities, swamps, and alpine zones. Some northern species hibernate. Most species breed in various kinds of decaying plant and fungal material, including fruit, bark, slime fluxes, flowers, and mushrooms. A few species have switched to being parasites or predators. Many species can be attracted to baits of fermented bananas or mushrooms, but others are not attracted to any kind of baits. Males may congregate at patches of suitable breeding substrate to compete for the females, or form leks, conducting courtship in an area separate from breeding sites . . . .

The genus Drosophila as currently defined is paraphyletic (see below) and contains 1450 described species,[3][4] while the estimated total number of species is estimated at thousands.[5] The majority of the species are members of two subgenera: Drosophila (~1,100 species) and Sophophora (including D. (S.) melanogaster; ~330 species). The Hawaiian species of Drosophila (estimated to be more than 500, with ~380 species described) are sometimes recognized as a separate genus or subgenus, Idiomyia,[3] but this is not widely accepted. About 250 species are part of the genus Scaptomyza, which arose from the Hawaiian Drosophila and later re-colonized continental areas . . . .

Drosophila are extensively used as a model organism in genetics (including population genetics), cell-biology, biochemistry, and especially developmental biology. Therefore, extensive efforts are made to sequence drosphilid genomes . . . . The data will be used for many purposes, including evolutionary genome comparisons. D. simulans and D. sechellia are sister species, and provide viable offspring when crossed, while D. melanogaster and D. simulans produce infertile hybrid offspring. The Drosophila genome is often compared with the genomes of more distantly related species such as the honeybee Apis mellifera or the mosquito Anopheles gambiae.

In short, the study reported by Zhou et al as linked in the OP, is based on a major focus on this genus, a genus that shows fairly wide diversity and apparently fairly high mutation rates. Doubtless, the projected timelines on the various members and evolutionary divergence are due to the various conventional dating schemes. (BTW, a Q: if “sister species” are interfertile, why are they regarded as separate species?)

The key inference in Zhou et al, as observed by PaV is: 6) the rate of the origin of new functional genes is estimated to be 5 to 11 genes per million years in the D. melanogaster subgroup.

That has led to the following calc in the OP:

Noting that Drosophila melanogaster has 14,000 genes (a very low gene number), the simply calculation is this: 14,000 genes/8 [i.e. mid point of the range 5 - 11] new functional genes per million years= 1.75 billion years for the formation of the fly genome. This, of course, assumes that somehow the fly is ‘alive, and reproducing’ the entire 1.75 billion years—-this, without the aid of a full-blown genome. If we apply this to the monkey/human difference which, IIRC, is about a 1000 genes, then using this same rate, it would take 200 million years for man to have evolved from the monkey. This published rate for new functional gene generation cannot be good news for Darwinists.

Obviously, the calc is in large part meant to highlight the basic problem: accounting for body-plan level genetic information within the time claimed to be available on the usual reported timeline for life on earth. That problem is of course most notoriously seen in the Cambrian life revolution, where dozens of novel body plans appear in a window reported as being about 10 MY, some 500 - 600 MYA.

600 or so MY is a lot less than 1.75 BY, i.e. something is wrong with rates and/or dates. Independently of dates issues, we know that the genomes and associated proteins are long-chain discrete state systems, i.e. information-rich, functionally specified polymers. This is already so problematic for the macroevolution and OOL models that we can focus on just this part, taking the usual dates as givens for this discussion.

The basic information storage capacity of a digital entity such as D/RNA or proteins, is reasonably estimated at 4^N for the former, and 20^M for the latter; and whatever variations in frequencies of component monomers we may bring up, this will not shift the basic order of the calculation. There just is not enough time, mass and material on earth — or in the observed cosmos — to get to the observed patterns and complexity, with reasonable probability relative to chance + necessity without intelligence. But, intelligence easily and routinely produces functionally specified information-rich systems that are of this order of complexity.

So, Q: why is there such a strong resistance to the idea that intelligence is a reasonable — indeed the best — explanation for the FSCI in life forms?

Answer: worldview level commitments, often backed up by attempted redefinitions of the nature of science that seek to constrain scientific inference to themes friendly to materialism.

So, there is a serious issue that needs to be seriously addressed.

GEM of TKI

PS: I have looked up a relevant timeline, and find one here:

The Diptera are commonly known as (true) flies and include many familiar insects such as mosquitoes, black flies, midges, fruit flies, blow flies and house flies. Flies are generally common and can be found all over the world except Antarctica. Many species are particularly important as vectors of disease in man, other animals, and plants. In addition, much of our knowledge of animal