Uncommon Descent Serving The Intelligent Design Community

Evolution Theory Fails? “It can’t be true!”

arroba Email

“To knock out 2 megabases and not have an effect—that’s remarkable,” says Jim Hudson, a geneticist at Open Biosystems in Huntsville, Alabama. “It can’t be true,” says a skeptical Arend Sidow of Stanford University.

Arend Sidow is not a dumb or unqualified guy. He’s the principal researcher at Stanford University’s Sidow Lab. In fact this is his lab’s primary area of investigation: Much of our work rests on a simple, fundamental, principle of evolution: functionally or structurally important sites in the genome are subject to selective constraints;

So why can’t it be true? Because if evolutionary theory (natural selection) is true then it MUST BE TRUE that conserved (constrained) regions of DNA have biological activity. Arend won’t entertain natural selection at the DNA sequence level not being true. The knockout experiment, if there is no mistake, overturns his faith in evolutionary theory. So the experiment must be flawed. “It just can’t be true.” Arend is having a crisis of faith. Isn’t that just precious?

THE BIOLOGY OF GENOMES MEETING: Disposable DNA Puzzles Researchers
Elizabeth Pennisi
11 JUNE 2004 VOL 304 SCIENCE

For a long time, any DNA that didn’t make up genes was considered junk, even though it constituted the bulk of the human genome.

Gradually, though, genome biologists have found gems among this non–protein-coding sequence, suggesting that “junk” was a serious misnomer. But new research suggests that vast tracts of this sequence may be disposable after all: Marcelo Nóbrega, a geneticist at Lawrence Berkeley National Laboratory (LBNL) in Berkeley, California, finds that mice can do just fine with millions of these bases deleted from their genomes.

About 2 years ago, Edward Rubin, director of the Department of Energy Joint Genome Institute in nearby Walnut Creek, and his colleagues discovered that some DNA sequence in human gene deserts—long stretches of DNA between genes—was almost the same as the sequence in comparable mouse deserts (Science, 31 May 2002, p. 1601). This conservation across species that shared a common ancestor more than 80 million years ago seemed highly unlikely—unless those regions served a purpose. Since then, “we’ve had the assumption that all these regions are doing something,” says Michael Zody, a computational biologist at the Broad Institute in Cambridge, Massachusetts; that something is probably gene regulation, he says.

Nóbrega and his colleagues found support for this idea when they compared desert regions conserved between fish and humans, quite distant relations. “Most sequences conserved between fish and humans are not only functional, but they help regulate genes,” Nóbrega reported. But their mouse-human comparison told a different story. They analyzed 15 comparable desert sequences and found that only one was a regulatory region in both species.

Puzzled, Nóbrega, Rubin, and their colleagues decided to delete deserts from mouse genomes, hoping to learn what other function these regions might serve. His LBNL colleague, geneticist Yiwen Zhu, knocked out two regions, one about 2 million bases long and the other about 1 million, both of which were conserved in humans and mice but not fish. After inserting the altered genome into embryonic mouse stem cells, Zhu added the cells to mouse embryos and looked for abnormalities in their descendents. “There was no sign of any difference in survival” between the genetically altered and normal mice, Nóbrega reported at the meeting. “There’s no sign of overt pathology.”

Other researchers are dumbfounded. “To knock out 2 megabases and not have an effect—that’s remarkable,” says Jim Hudson, a geneticist at Open Biosystems in Huntsville, Alabama. “It can’t be true,” says a skeptical Arend Sidow of Stanford University. Both Hudson and Sidow wonder whether these noncoding regions have a function that just doesn’t show up in the tests Nóbrega did. So the question remains open, says Rubin: “Is the genome like a trash novel, from which you can remove 100 pages and it doesn’t matter, or is it like a Hemingway, where if you remove a page, the story line is lost?”

Is there something about the spam filter on this blog that I need to know about? My last comment may have been a bit wordy, but I didn't think it was particularly provocative. Indeed, it sort of leads down the "systems" approach road, if not exactly as everyone here expects. Just wondering. Art2
I think it might be relevant here to suggest another way of looking at the issue. Is there a simple quantitative information theoretic way of showing that much of the supposedly non protein-coding "junk" DNA must still be functional? What I have in mind here is the total information content required to specify the structure of an organism combined with its phylogenetic development schemes from egg and sperm. If the genome must contain all of this information, a minimum number of nucleotides would be needed to encode it. For humans the protein-coding portion is supposedly only 2% or so of a total of 3 billion (60 million nucleotides). To be conservative, doubling that would get to around 120 million "letters" in the genetic alphabet and 5 million specified amino acid "words". This of course grossly oversimplified, but seems to be reasonable within maybe an order of magnitude. Following this approach, 5 million "words" would be equivalent to about 40 books of 400 pages with 40 lines per page and 8 numbers per line. It seems clear to me at least that the protein coding portion of the genome is grossly inadequate to contain the necessary information. The 40 textbooks would barely be able to document the structures and developmental programs for the individual cell types themselves - the multiple coordinated molecular machines like the nucleus and organelles like mitochondria, ribosomes, Golgi apparatus, etc. etc. Even this must be infinitesimal compared to the informational equivalent of the body and brain including developmental programs. It must be truly astronomically huge. There are 10 trillion cells total. The brain is supposed to be the most complicated single object in the universe, containing 100 billion neurons each of which is connected to several thousand others in a very ordered structure. Thus a very rough estimate of information content seems to indicate the need to utilize vastly more than the normally considered protein coding part of the genome, in order to specify the total physical design and development programs. This reasoning is just covering the DNA that in principle is needed to build and operate the organism. The point seems even clearer if the genome has to contain "spare parts" or evolutionary toolkits and other nonfunctional in life but still needed DNA not needed to build and operate the organism. This could include "front-loaded" genetic information intended for future use. This reasoning strongly implies that much of the "junk" DNA subject to the ENCODE research must somehow be part of this information storage. This would include the apparently conserved but nonfunctional regions discussed in this thread. magnan
bFast. If it isn't too much trouble. Could you please provide a quick summary of your views? Thanks. christopheratlee
Great_ape, "I don’t think it’s not being done because of the evolutionary paradigm, however. I think it would be hard to convince a funding agency to go for it…" I think its hard to convince funding agencies because they don't want to go the way of Guillermo Gonzalez. Sorry for not having faith in the impartiality of the scientific community -- but recent, clear evidence is on my side. bFast
Edward Rubin says:
Two important issues, however, limit the use of this ‘middle of the strait’ approach to identifying functional elements in the human genome. First, some functional elements are certainly either human- or primate-specific, and accordingly, will be missing from the genomes of non-primates9. Second, the enormous degree of noncoding sequence conservation that is found between humans and mice is probably the consequence of nonuniform rates of evolution across the human genome10. This results in genomic sequences that still show a considerable degree of similarity that reflects a slower evolutionary rate rather than purifying selection11,12.
Translated: Purifying selection is dogma so when it fails a prediction we don't question the dogma we just crank in a new epicycle (different rates of purifying selection in different parts of the genome) to account for it and continue merrily upon our way with our faith in natural selection being the center of the biological universe intact. To speak otherwise, as Galileo used his observation of the the moons of Jupiter to support heliocentrism, is to incur the wrath of the Grand Inquisitors of the Darwinian Church. Fortunately you don't get burned at the stake anymore for questioning the sacred screeds but blackballing and denial of tenure are popular substitutes. Sarcasm aside what Rubin is really saying is that some non-functional DNA for some unknown reason is maintained so well that over the course of 100 million years of purifying selection you still can't tell it apart from functional DNA. That makes no sense at all. Why would it possibly be the case under the rubric of natural selection that non-functional DNA would be so immune from purifying selection that it remains as intact as functional DNA? Nobody seems to be asking that question. That's because, as I and others have said, Darwinists stopped testing the theory of evolution a long time ago. I think a big reason for this situation is that testing it doesn't pay dividends. There is no practical application for macroevolution (evolution that takes place over millions of years versus scores of years) in modern biology. Large scale evolution works too slowly (if it works at all) to have any practical consequence in timeframes that people need be concerned about. DaveScot
My understanding is that this experiment is a few years old already. By now we should be getting results from further large-scale experimentation. --bFast I would certainly hope--and rather suspect--people are following up on these and similar large-scale genomic knockouts. There may have been a follow-up published that has flown under my radar; I haven't looked into it specifically. The logistics and expense of conducting the sort of experiment you suggest (are something equivalent) are prohibitive. There is also the matter of just how long it would take for the benefit of the removed region to manifest itself in terms of increase/decrease in population size. But, being a curious type, I'd like to see something vaguely along these lines conducted. I suspect others would as well. I don't think it's not being done because of the evolutionary paradigm, however. I think it would be hard to convince a funding agency to go for it... great_ape
bFast It appears everyone doing practical research (that excludes evolutionary biologists) expressed some shock that 100my separating mice and men isn't long enough for neutral drift and natural selection to reliably separate the wheat from the chaff in biologic activity of DNA sequences and have moved on in search of some other non-experimental way of reliable discrimination. It's really important for medical progress that our most important model animal for experimental use (mice) be better characterized. Comparing fish and human genomes to tease out active conserved sequences is far less useful than being able to do it with mice. Researchers trying to find potential therapeutic drugs for humans don't test them on fish. They test them on mice. And unfortunately even such a much more closely related organism like the mouse is very often not close enough to reliably predict any given drug will have the same effect on a human as it does on a mouse. At the present time it seems the only reliable way of discriminating between active and inactive sequences shared between mice and men is to knockout the sequence in the mouse and see what happens when you raise the GM animal which is a long, laborious process compared to what would have been a panacea had Rubin's deletion of 1000 highly conserved non-coding sequences from the mouse revealed that conservation=activity is a falsehood. DaveScot
Art You weren't difficult to find. Biochemistry PhD, 1982, Brandeis, correct? So should I believe you or Behe since you seem to have more or less equivalent credentials? I'm still going with Sidow. His remark isn't all that dated. It was made barely 3 years ago. The only thing that's happened since then is molecular clocks have come under even more disrepute. Calibrating one is like calculating orbits using epicycles. Furthermore, Edward Rubin runs the Genomics Lab at Lawrence Berkeley National Laboratory, he expressed amazement (again barely 3 years ago) and he certainly has far better credentials than you. I don't really like playing the credential card but when it's this far outside my expertise I have to rely on others and the only way I have to assess their credibility is through their accomplishments and then take them at their word. Sidow, Rubin, and others were all shocked at the result and quoted in reliable trade rags under their real names. You appear to be an outlier and express your lack of shock anonymously on a blog. If you think you've got the answer to this puzzle this isn't the place to present it. Write up a review article and submit it to Cell or Nature or some equivalent peer reviewed source. If it passes muster there I'd be more inclined to seriously consider it. DaveScot
Great_ape, "Is there some specific aspect of biology being neglected that would bolster ID in your mind?" Yes -- there are two discussions in this thread discussing conserved DNA that shows no apparent function when knocked out. A number of biologists are recognizing that the only mechanism known to MET to conserve DNA in this way is that the DNA has function. Lets test the thing. Lets get a thousand mouses, half of which have the knocked out gene, set up a containment (say an island) and let 'em live a more real life for a few years. If there really is no deleterious effects in the knock-out mice, the ratio should remain about constant. If the knock-out mice diminish, then MET is vindicated. What if the knock-out mice actually begin to dominate? My understanding is that this experiment is a few years old already. By now we should be getting results from further large-scale experimentation. All of the above also goes for ultra-conserved segments that can be knocked out without obvious effects to the organism. bFast
Has anyone done any snooping to see if there might be any super secret messages encoded therein? --mike1962 The short answer is yes. After several months of painstaking effort, computational biologists were able to decode the following brief missive: "DRINK MORE OVALTINE." Scientifically, we're not quite sure how to proceed with this new information, but, on a more personal level, we're all still trying to cope with the far-reaching implications of this discovery. great_ape
The link didn't work in comment #7. Here it is: http://genome.ucsc.edu/cgi-bin/hgGateway?hgsid=82408863&clade=vertebrate&org=Mouse&db=0 Sorry about that. Art2
I'm not certain I follow your argument, bFast. As far as I can discern, all manner of biological reality is being investigated, given a sizable but not limitless, amount of resources. Is there some specific aspect of biology being neglected that would bolster ID in your mind? In my opinion, researchers are generally trying to learn as much as they can about as many things as they can. If there are facts out there that support ID, they may come out in the wash during this process. However, I personally can't think of any fruitful ID-centric research programs that don't significantly overlap subjects already covered by the evolutionary paradigm. Remember evolutionists studying the "illusion of design" care every bit as much about how and why an apparent design works as does someone who thinks it was, quite literally, designed. great_ape
Has anyone done any snooping to see if there might be any super secret messages encoded therein? (I guess not, or we'd know about them.) Maybe it's time somebody looked. Yeah, I know I'm a kook, but you never, never know. mike1962
great_ape, "These are possibilities that need to be carefully investigated before I would accept more radical hypotheses." This statement is fine enough, yet the scientific community spends millions, nay billions, to validate the big bang theory. Isn't that a prime purpose of the hubble telescope? It seems to me, however, that the biological community has said, "we know our theory is right". It then says, "these possibilities need to be carefully investigated ...", it then is uninterested in putting out the energy/money to do the research. After all such research could be construed to be ID research, which is taboo. Further, we know our theory is correct so we don't need to spend billions, nay millions, to validate it in light of challenges. bFast
Hi DaveScot, You said in the other thread (is that what they call these things on blogs?):
Bona fide researchers in this area of expertise such as Arend Sidow disagree with you. What can I say? Should I put more stock in an anonymous blog commenter or people like Arend Sidow? No disrespect meant but I’m going with Sidow’s opinion over yours. See my recent article (I wrote it with you in mind) quoting Sidow’s reaction to the experiment which you don’t think means anything.
Thanks for the thoughts. It's nice to see that someone is reading my comments. Listen to Sidow you should. But I think I am speaking with several more years of data to work with than he had when he made his remarks. And while I may be anonymous, I believe at least one UD contributor (who will remain anonymous, so as to spare embarrassment) will vouch for me. I’ll try to explain my ideas and then bow out, because I don’t think either of us is interested in an extended “he said, she said” back and forth. (I’ll answer questions directed to me, of course.) To appreciate what I am saying, one really should go to this web page and choose the latest human genome assembly, using the appropriate pull-down menus at the top. Type in the corresponding human genome position of one of the gene deserts (it took awhile for me to get this set of coordinates, because things have changed since the original paper – I know it’s correct because I found one of the genes mentioned in the paper): chr1:82,073,954-84,900,793 What will come up is a screen with several lines of display. The first thing to keep in mind is that, at the time the paper in question was published, this “desert” extended from the LPHN2 gene on the left to the PRKABC gene on the right. Today, we can see that half of this “desert” actually consists of predicted genes (including spliced ESTs). So half of the “conserved” patch that was considered to be non-functional really isn’t. Now, look at the bottom seven lines of the picture – this illustrates the conservation between humans and mouse, rats, dogs, ‘possums, chickens, frogs, and fish, respectively. A quick glance shows that the conservation is pretty much nil at the bottom, and substantial on the top, as expected. I don’t think the original authors had access to most of these “lines”, but rather were focused on conservation between mice and humans. Does this gene desert actually have unexpected regions of high conservation? By unexpected, I mean things that we could not see resulting from random drift (which, I emphasize, is what the line-by-line comparison shows). To understand why I say no, one must look at the sequence level. One can click a few links and get there – I’ll just summarize what one will find. Basically, the vast majority of the differences between mice and humans (as well as most of the other differences in the mammals, at least) in this gene desert consist of insertions and deletions (indels), not point mutations. Random indel creation is expected to produce stretches of high divergence, and of high conservation. At the outset, there are lots of conserved stretches. As time passes, the conservation diminishes, both in terms of the size of the conserved patches, and in terms of sequence identity. It eventually disappears. This is exactly what we see in this gene desert. This is, I suspect, why later comments by the author (as you noted in the other thread) indicate that the conserved regions here are happenstance, nothing more. (One can get a feel for this by modeling the random insertion process using random number generators; you can even do this in Excel.) That is, in a nutshell, why I am saying that this result is really not that unexpected (in spite of the remarks in the original paper). Moreover, I suspect that today Sidow would agree with me. For what it’s worth. Art2
If I may add, there is great difficulty mathematically justifying that such a large number of designs can even be maintained by selection. There simply may not be enough population resources to police billions of nucleotides. See: Nachman's U-Pardox By the way, Nachman's U-Paradox is a complication on top of Haldane's Dilemma. There is institutional insentive, unfortunately, to try to keep arguing these mega-bases have no function (as this would tend to destroy both Darwinian and Neutralist theories). That is unfortunate. DNA researcher Andras Pellionisz has lamented that this institutional bias has stymied medical advancement. scordova
“no phenotype” is a term I personally take with a grain of salt. There are some genes that, if removed, show “no phenotype” until the organism is infected with a certain variety of bacteria.
Yup. These are contingency designs which may not be immediately visible to selection. You gave a very good example right there. scordova
I still find these megabase-level knockout experiments intriguing and perplexing. From a skeptical perspective, though, "no phenotype" is a term I personally take with a grain of salt. There are some genes that, if removed, show "no phenotype" until the organism is infected with a certain variety of bacteria. Then those without the gene die much more readily than wild type organisms. There are also extremely subtle behavioral differences that may go unnoticed if the organism has not been placed in the appropriate "natural" context (Which usually doesn't consist of cages, water bottles, and pellet food.) These are possibilities that need to be carefully investigated before I would accept more radical hypotheses about genome structure and organization. great_ape
One of the advanced ID theories, which I will call Biotic Steganography, was advocated by William Dembski. If one accepts the ID hypothesis of a Privileged Planet, which argues that nature is designed for scientific discovery, then Biotic Steganography is an extension of that notion in Biology, "Biology is designed for scientific discovery". The conserved regions are there so that we may understand ourselves. In my heart, I believe that to understand humans, we must study all the creatures which the Designer has made. He ordained that for us to understand ourselves we must also study all biology and all nature. It was by design. The conserved sequences are no accident, but constitute the Rosetta stone from which we can discover all sorts of things about ourselves. It is no accident we have so much similarity to the Apes. Apes are a gift from the Designer to help us understand ourselves. The ape genome will give us clues to assist us in making medical advancements. The day will come when we will say, "it's a miracle apes exists, were it not for their genomes, we would not have found a cure for this disease!". I wrote on Biotic Steganography here: How IDers can win the war. That is why ID is a better vision for medical and technological advancement than Darwinism. scordova
I had a major chat about this on Telic Thoughts a while back. Not only does the mouse experiment exist, but there are at least 9 ultra-conserved* non-coding regions which show no obvious deleterious effects when knocked out. I personally believe that MET cannot withstand even one such region. *Ultra-conserved, if I understand, means within 98% identity in all vertibrates. bFast
Much of our work rests on a simple, fundamental, principle of evolution: functionally or structurally important sites in the genome are subject to selective constraints;
I have pointed out in the past where this reasoning is completely flawed. Complex systems entail redundancy and capacity for self-healing and self-correction. These capacities can be mathematically and empirically shown to be almost if not completely invisible to selection. See: Airplane magnetos, contingency designs, and reasons ID will prevail. The failure of knockout experiments to detect design has been anticipated by numerous ID proponents.
“This knock out approach misses the fact that there are alternative genetic routes, or pathways, to the production of the same cellular product. When you knock out one gene, the genome can compensate by using an alternative gene. But when you repeat the knock out experiment by deleting the alternative, the genome can revert to the original gene instead. Using the knock-out approach you could infer that both genes are expendable from the genome because there appears to be no deleterious effect in both experiments.” Using the knock-out approach you could infer that both genes are expendable from the genome because there appears to be no deleterious effect in both experiments.”
I said it last year, and I'm glad to seem my hypothesis is getting vindication. The way we can detect such designs which are invisible to selection is through application of the Explantory Filter. Biologists are using antiquated detection paradigms because of the Darwinist insistence that functionality needs to be selective. That Darwinist hypothesis can be mathematically and empirically refuted, and it's evident by these recent developments in peer-review. scordova

Leave a Reply