[Here is something I just sent Casey Luskin and friends regarding the ENCODE 2015 conference. Some editorial changes to protect the guilty…]
One thing the ENCODE consortium drove home is that DNA acts like a Dynamic Random Access memory for methylation marks. That is to say, even though the DNA sequence isn’t changed, like computer RAM which isn’t physically removed, it’s electronic state can be modified. The repetitive DNA acts like physical hardware so even if the repetitive sequences aren’t changed, they can still act as memory storage devices for regulatory information. ENCODE collects huge amounts of data on methylation marks during various stages of the cell. This is like trying to take a few snapshots of a computer memory to figure out how Windows 8 works. The complexity of the task is beyond description.
To get a hint of the importance of the methylation marks see:
http://www.nature.com/nature/2015/180215/full/nature14310.html
So it’s just my conjecture, the repetitive DNA is there to act as dynamic store of data. We are fooled into thinking since the sequences don’t change much during an organism’s lifetime that it can’t act as an information processing system since we presume the relevant information is the DNA sequence, when in fact the critical information are the methylation marks.
The repetition is probably important for the organism to recognized the regions and say, “hey, here’s where a lot of RAM is for me to use.” The mistake is thinking the significant information content is in the “ACTG” sequences, it is not, it is in the methylation markings, and that could well be where some serious parts of ontogenic regulatory information is flowing through.
Gruar’s approach is akin to opening up a computer and examining the physical RAM in it and saying, “those VLSI transistors are all identical, look at all that repetition, therefore it’s junk!” What matters is when the computer is up and running and we’re seeing these transistors switching back and forth from 0 to 1. Those zero’s and ones are the DNA methylation marks for regulation (not the repetitive ACTG transistors), and that’s why I suspect Graur is way off mark. It really is too early to tell, I wouldn’t yet go out on limb, but Graur is going out on a limb, and if proven wrong, we can publicly call him on it.
Repetitive sequences are plug and play like computer RAM, and can be subject to some variation. But they need to be there.
This conjecture seems wholly divorced from any known biological mechanisms. There is no analogue in biology for an operating system’s ability to access arbitrary sites in memory by address; all regulatory systems involving DNA ultimately depend on features of the DNA sequence itself. Quite a lot is known about many regulatory pathways, including developmental pathways, and none of them involve the kind of mechanism proposed here. Your suggestion appears to invoke an unknown process to explain nothing in particular.
Moreover, under this proposal, the specific sequence of repetitive DNA is critically important, since it only functions if it’s identical to other elements. But most repetitive DNA shows no signs of selective constraint, as would be the case if this were true.
I don’t see any justification for taking this idea seriously.
Steve,
You don’t think methylation marking and positioning of the is important to regulation? I don’t think you’re up-to-date on the latest. You’re still stuck reading evolutionary biological speculations, not looking at molecular biological evidence.
Good gravy, I just linked to a February 2015 paper that suggests otherwise. Regulation depends on the epigenetic state, not just the DNA.
That’s why the 213-215 different cell types which mostly have about the same DNA have different regulatory implementations. Cancer researchers are acutely aware of this, and they are a large part of the ENCODE consortium.
stcordova,
Trace back any epigenetic state, you get to some other DNA. DNA is a substrate for, as well as the source of, DNA-sourced ‘code’.
Ahh, more ENCODE announcements on 40 year old stuff. DNA is methylated when accessed? STOP THE PRESS!
Just curious Sal. Are you working towards some number goal of threads started and then bailed out on when the questions got too tough? 10? 20?
Of course I think it’s important. What I said was that methylation, like other regulatory systems, ultimately depends on DNA sequence. DNA methylation occurs at specific sites because of specific DNA sequence motifs; for example, see this paper on how the transcription factor Myc establishes methylation patterns by recruiting a methyltransferase. Transcription factors work by binding to specific DNA motifs, not to arbitrarily chosen and interchangeable pieces of DNA.
There is no known mechanism for picking out randomly chosen repeats for methylation and then using the results for regulating specific genes. Real epigenetic modification regulates a gene by interacting mechanistically with the DNA needed to transcribe that gene, not by carrying out computation in an arbitrary scratch space of repeats.
Why you’re appealing to ENCODE results is not clear to me, since their studies find epigenetic modifications that happen consistently at the same sequence location for particular cell types. That approach would never find the kind of RAM-like behavior you’re suggesting.
I’m not a molecular biologist and gene regulation is not my field, but I’m pretty sure I know more about it than you do.
It doesn’t completely depend on DNA, DNA is a necessary but not sufficient system to implement regulation. The central dogma could still be true in the sense DNA isn’t usually re-written by the epiginome, but it doesn’t mean all the regulatory technology for things like ontogeny is stored in the DNA. Just having all the proteins coded and in the right proportions don’t make a frog — just put one in the blender and you’ll see what I mean. 🙂
Actually, don’t we now know that the central dogma is in fact false?
It’s more than that, didn’t you even look at the 2015 paper in Nature.
But not ALL the time! The epigenetic state changes the methylation marks. Put some of the mammary gland DNA from Dolly the sheep in a mammary her mammary gland cell, and it stays a mammary gland cell. Put that same DNA in another epigenetic context and you clone a sheep!
Ergo, not all the technology to build a living system is contained in the DNA, otherwise you could just lay the DNA in a bottle and it will replicate in to cells. Just because the epigenome rarely re-writes the DNA doesn’t mean DNA contains all the technology to implement a cell.
What makes you think DNA contains all the technological capability to create cells? Can you tell the readers where the locations of ontogenic information for multicellular Eukaryotes? If you can’t, you’re being premature at best, and you’re dead wrong at worst to say DNA contains all the regulatory and ontogenic information.
Congrats on being picked up by UD, Sal.
No, why?
I said:
Steve Shafner said:
Methylation marks are part of epigenetic memory, the prestigious scientific journal Nature’s website says:
http://www.nature.com/subjects/epigenetic-memory
So, repetitive sequences (which are made of DNA) can host epigenetic memory! So which idea is wholly divorced from what we know about biology. Not mine.
I once suggested in so many words that I might just obliquely state what’s in textbooks, scientific papers, and conference journals — and somehow my detractors will find a way to say I’m totally wrong. Then all I have to do is deliver references later showing I articulated a mainstream view, and then who will look a little awkward?
Repetitive sequences can serve as a host for epigenetic memory, and thus have regulatory FUNCTION. They are not necessarily junk just because they are repetitive sequences any more than a repetitive array of VLSI transistors are junk just because they repeat! The ATCG DNA letters are not the only information that is being stored, these sequences store epigenetic memory.
Graur’s approach is like looking at a repetitive array of of VLSI transistors on a computer chip and saying they are junk simply because they are repetitive.
Nonsense.
Did Dan Graur even mention “epigenetic memory” or methylation marks in his screed:
http://gbe.oxfordjournals.org/content/early/2013/02/20/gbe.evt028
No! He’s just whining, he’s not dealing with accepted science.
LINE-1’s are repetitive elements. I said in The Sugar Code and other -omics
And lo and behold, what do we find that a repetitive element such as the L1 SINE is able to do, bwaha!
http://www.annualreviews.org/doi/abs/10.1146/annurev-cellbio-101011-155822?journalCode=cellbio
How does this repetitive element do it? Remember it’s not just the DNA sequence it’s the methylation markings, baby!
Literature abounds now that DNA carries information beyond the ACTG.
Here is a something from wiki in regarding neuroscience and how DNA is possibly involved in your ability to remember. Though the DNA genetic sequence in your brain cells are pretty much the same in each cell, their epigenetic marks on the DNA are different for each cell, and this may be a part of how you remember things!
https://en.wikipedia.org/wiki/Epigenetics_in_learning_and_memory
I said in the OP:
Maybe I understated my case by mentioning only regulatory information, maybe even some of the brain power to remember.
From the wiki link above
Mung,
No, we don’t.. It is frequently misunderstood, even by biologists, but its original statement (and Weissman’s prior ‘doctrine’) remain unfalsified. Sequential information flows from nucleic acids to proteins, never the other way round. Show me an exception and I’ll tell you why you’re wrong. 🙂
stcordova,
All epigenetic markers originate from DNA. They may not originate with that particular copy of the DNA molecule, but they do originate from that particular genome – from a parental cell, or one with which that particular molecule shares common ancestry.
I don’t know wh you are getting so excited about methylation – it is little different from the lac operon, discovered in 1961 by Jacob and Monod. DNA-level control mechanisms affect transcription, by several different potential pathways. Not every gene is active at once, and no-one ever thought they were.
stcordova,
The article is behind a paywall. You didn’t just read the abstract did you?
eta – but anyway, ‘how it does it’ does not appear, from the abstract, to be related to methylation, but rather the consequences of retroposition itself.
If you are looking for ways that memory might be digitally encoded, then count me as skeptical.
It wouldn’t matter if it was (though I agree it’s not a strong contender). There’s nothing about digitalness in itself implies a thing is designed.
And in any case, firing neurons are kind of digitalish.
I didn’t just read the abstract, that quote about L1 and neurons having separate transcriptomes was toward the end of the article.
I have university access to paywalled articles. Since Dan Graur and Larry Moran are at universities, they can’t make the excuse of ignorance due to paywalls for the dumb stuff they say, maybe ignorance due to their insistence since something repeats and jumps around the geneome (like the L1 transposon) that it must be junk…..
FYI, some of universities as a public service (if they are publicly funded) allow public access if you actually go to their libraries (rather than just online “visits”) to paywalled articles.
Sal, if Larry is so dumb, hike over to his site and confront him. It’s a current topic.
HAHAHAHAHAHAHA.
stcordova,
Larry is very clear that he does not regard active transposons as junk. Broken transposons, yes. They no longer have any dynamic affect on gene regulation either, when they are broken, so they don’t form part of your ‘epigenetic army’.
Larry adds up all DNA having any kind of function, and comes up with about ten percent. I still get the giggles thinking about Sal going head to head against Larry.
Thank you, Rich.
Am still hoping we’ll get some mention on the Discovery Institute website. 🙂
So how does Larry know a given transposon was never active! In an adult human there are 200 trillion cells, and through development, even more — to establish they were never active, he’d have to look at every cell at every developmental stage, and we know some transposons are only active at specific stages of development.
Looks like Larry’s going to lose this one because they just keep finding more and more — that paper is one example.
Larry bases his notion that it’s junk on his notions of Evolutionary Theory, not actual experiment and observation. An otherwise brilliant Princeton Chemistry PhD says dumb stuff because of his commitment to his interpretation of evolutionary theory which has little in the way of telling about the functioning of the cell. If one wants to make claims about the function of the cell I’d think molecular and medical biology are more informative disciplines than concocting stories from phylogenetic trees and then using those concocted stories to determine the function and physiology of a cell.
He’s adding up with incomplete accounting, we haven’t even scratched the surface! 200 trillion cells, and even assuming 5 stages of development for each, that’s a quadrillion. And the modest ability to self-heal — there is a lot of latent capacity that might not be realized unless contingencies are activated. He’s premature at best, dead wrong at worst.
He’s trying to persuade other scientists these DNA regions are non-functional without even looking? That’s religious belief and evangelism with no facts. We’ll know better AFTER looking, and that’s what ENCODE is about.That’s science, and that’s not what Larry is doing here.
He should know better, he’s a chemist where experiment and observation count for something, and he’s abandoning all of that to make pronouncements with highly incomplete data sets.
ENCODE and other projects are going to squish Dan Gruar and Larry Moran.
Not only is there the 300 or so million dollar ENCODE project, but there is now the 200 million dollar ROADMAP project to boot, it studies epigenetic memory:
http://www.roadmapepigenomics.org/
and overview
Here are some basics, courtesy of the United States Government:
http://www.genome.gov/27532724
Everything I said in the OP is supported by these basics.
Sal, if you have questions for Larry, go to his site and ask. We’ll watch.
Someone will respond, and if the discussion is interesting, Larry will join in.
I’ve gotten the responses from om him.
The only novel thing in your OP — your proposal that repetitive DNA acts like RAM — is completely unsupported by these basics. As I pointed out previously, it’s actually contradicted by them, since every result from ENCODE and from the Roadmap is based on finding the same epigenetic changes at the same loci in the same type of cell. Basically, your proposal runs counter to everything that has been learned about genetic regulation in the last 50 years.
I went to his site. When I signed in to wordpress it associated me with one of my blogs, “LiarsForDarwin”. That’s me.
I mentioned the L1 transposon is involved in creating separate transcriptomes in every neuron cell. L1 is an example of what Larry calls, “junk”. Why? Because it’s repetitive. Using the same stupid logic one could argue that “since the VSLI transistor arrays are highly repetitive, they must be junk”. That’s Larry logic, not science.
This just in:
http://darwins-god.blogspot.com/2015/07/nih-director-each-neuron-is-different.html
Larry loses another round.
Steve Shaffner,
So are you saying DNA doesn’t contain unique epigenetic information in each cells epigenetic memory?
RAM has the same loci in a computer over time, and the same loci in the replicas of that computer, otherwise it can’t be addressed! But the contents of RAM can change.
The issue isn’t the loci of the methylation marks, it’s differing methylation marks at the same loci (where by loci I mean a stretch of DNA). The methylation marks change during stage of the cell.
This isn’t my proposal, it’s there in the literature.
Here is another link on DNA Methylation. Get a load of them mice that have the same DNA but different methylation patterns!
http://www.ks.uiuc.edu/Research/methylation/
What’s that Larry? You a biochemist know about information that can be encoded in DNA which actually makes repetitive DNA sequences actually informative, and you’re not acknowledging it. Shame on you.
Just curious as to whether or not Sal still believes that an 8-nucleotide stretch of DNA accumulating one new mutation per “generation” per site and thus in 8 “generations” is totally different from the starting sequence is not only a realistic model but a refutation of evolution as he did about 10 years ago.
Sal,
Please explain how methylation increases information, and how this is quantified, and how this supports ID creationism.
stcordova,
I don’t think Larry thinks that there is any such thing as a ‘transposon that was never active’. That would just be bizarre. What’s happened to you. Sal?
stcordova,
No. The genome can be split into LINEs, SINEs, etc etc etc, and active ones can be distinguished from inactive ones. He did a detailed breakdown some time back, though I can’t be arsed looking for it as I know it will make no difference to you. You think a tiny trickle of new functions will somehow multiply up into 100% or so. Dream on. Get anywhere with those limestone calculations?
stcordova,
Go have a debate with him.
stcordova,
I just told you that Larry does NOT call currently active retrotransposons junk. You even responded to that, kinda. And here you are repeating it. Do you expect to be taken seriously?
L1 transposons are 17% of the human genome.
From wiki on transposons
Immobilization does not mean inactive, and by the way that wiki entry referred to an older paper than the one I provided on Neurons which indicated transposition events are usually not detected because they are cell specific, so the transposition event might just happen in 1 in trillions of cells, and if we’re not looking we’ll miss it!
Larry’s is premature to be making pronouncements.
If Larry wants an internet debate he’s on. He can come here or I can go there. I posted there, and for whatever reason my handle was “LiarsForDarwin” rather than stcordova. WordPress seemed to use the name of my blog rather than my username.
If L1s are 17% of the genome, and if we haven’t done proper accounting of all cells in all developmental stages, how can he say the genome is 90% junk. Surely he’s premature! And the only way we’ll have a chance of knowing is through work like ENCODE and ROADMAP which he is clearly against because of his ideology.
stcordova,
OK I’ll do the legwork for you
Go talk to Larry. He has no particular reason to come here, he has his own blog.
If a transposon is declared defective, this is because one of the essential genetic elements to allow it to transpose is damaged. Of course genomes are sequenced from readily available tissue, so it is possible that the element is only defective in that tissue. It would have to be competent in the zygote for this to be so, however, and if it were, one would expect to detect the consequences of that in the population. AFAIK, we don’t.
No. I’ve repeatedly said that I’m not saying that.
It’s not just that the contents of RAM can change; any slice of RAM can contain any piece of information, and the information works equally well whichever slice its in. Those are the unique features of computer RAM, and those are the features that are missing from epigenetics.
Yeah, I know. (Also, the singular of “loci” is “locus”j.
Allan,
Thank you for your comments and legwork.
I have to point something out, the numbers Moran gave are tentative at best. There are trillions of cells in the human at various developmental stages.
We sample only a small fraction of those cells to do studies, and many of those cells are immortalized cancer cells. The relationships are rarely deterministic and linear because of the deep redundancies in cell.
How do we discover suspected function? A researcher looks at Genome Wide Association Studies (GWAS). For example a GWAS study might be done on dialysis patients and we look at 1000 individuals and see correlation to non-coding sites. We then try to knock out those sites and shove them in zebrafish and see if the correlation indicates causality. Many times it is not 100% correlated, only partially so, but there is causal influence.
That was exactly one of the methods employed by this ENCODE affiliated researcher:
http://www.med.upenn.edu/apps/faculty/index.php/g275/p8487177
Many ENCODE researchers have similar lines discovery. Here is another ENCODE researcher who looked at Vitamin D receptors. He had a huge team of researcher to help him on identifying the few non-coding strands involved in vitamin D recptors:
http://iem.ucsd.edu/people/profiles/christopher-glass.html
All this to say, we don’t find the function until we look, and it takes millions of dollars just to elucidate a few small regions at a time.
Just taking a few cell lines in vitro and extrapolating you didn’t see function in a transposon is horribly insufficient. It is extremely difficult to detect the roles of non-coding sequences, and some GWAS studies in the future will involve millions of patients because the detectability is very difficult.
If we don’t look we have a very low probability of finding. What I object to is Larry saying it’s not worth it to even look. The medical community is willing to look because they wish to leave no stone unturned, and they are finding function and reporting it through ENCODE and ROADMAP.
ENCODE and ROADMAP will prevail because they are already influencing development and treatment of human suffering vs. sticking their head in the sand like Larry over in Sandwalk regarding non-coding DNA.
As with the L1’s in the neurons, it took some clever work to discover there was a lot of transposons active that we didn’t expect. The result was surprising. But if you don’t seek, you likely won’t find.
Larry’s accounting is premature, we don’t have all the data. And who is collecting the data? People like those in the ENCODE and ROADMAP consortiums.
Take a look at the number of researchers on the vitamin D receptor who found function in the non-coding regions associated with vitamin D:
http://cmm.ucsd.edu/lab_pages/glass/glasslab/index.html
That’s only one small stretch of non-coding DNA. It’s really premature to be making broad claims of no function. We don’t know for sure, and it’s premature to say either way, but ENCODE and ROADMAP suspects there is a lot of function and the only way we often find out is to be looking at the entire genome constantly and over many many individuals through GWAS studies and who knows what other procedures.
Sal, the main reason we can conclude that vast tracts of the genome are non-functional is that they accrue mutations without ill-effect.
if they were functional, mutations in those regions would have an effect.
That’s like saying the 5 backup navigation systems of the space shuttle are non-functional because we can delete all 5 of them with no ill effect.
How many contexts are examined to demonstrate no ill effect for human biology? Not many relative to all the possibilities.
The GWAS studies show there is ill effect only in a fraction of some cases, but above random chance. That indicates at some point the margin of safety is compromised by mutations in these regions. Some of the GWAS studies note effects of even single nucleotide effects, but it won’t appear in all patients, so it shows that a lot of things will have to go wrong before the compromise in function (or margin of safety) is compromised.
The ability to delete and have no ill effect could just as well indicate alternative regulatory pathways and deep redundancy in biological systems. That should be quite obvious because we can self-heal many injuries.
I pointed out some embarrassing errors that happened in knockout studies (which apply to mutations in supposedly non-functional regions):
http://www.uncommondescent.com/intelligent-design/airplane-magnetos-contingency-designs-and-reasons-id-will-prevail/
The difference is that those backup systems were designed to be backup systems explicitly. We know they are backup systems. You don’t know anything of the kind about the regions in question, do you?
They could just as easily be the designers shopping list. Why don’t you design an experiment that will determine the truth of the matter?
Glass houses etc.