The NY Times has a response to responses to Carl Zimmer’s article on junk DNA, and there I am, the first citation. One thing that’s interesting is that it mentions that there were mixed reactions from readers, with some pushback to the idea…but all the scientists they quote on the subject have no problem at all with the concept of junk in our genome. I looked hard for a link to someone levying a solid criticism, and it ain’t there.
Maybe they don’t exist?
Even Zimmer’s original article needed to strawman the pro-junk position, claiming that the pro-junk camp believed that all the functional DNA had already been discovered and further investigations are futile. That way it can all be tied up with a nice bow, saying “See! BOTH SIDES have something to learn from the other!”
Going back and reading Zimmer’s article, he explicitly says there is a controversy and cites Cell on Rinn’s work as a major scientific contribution to the debate. It’s unclear why he harked back to Gould and Lewontin who didn’t speak to genes directly, instead of Dawkins, who if I correctly remember the way he parsed it, claims to have demonstrated that junk DNA is an inevitable effect of gene competition. The names Gould and Lewontin savor respectively of dead and fallacious perhaps?
I don’t think there’s really much doubt that Zimmer is squarely in the anti-junk brigade. A while back he did a fairly sizable piece for the BBC on major blood groups, covering the mystery of how they are adaptive…without ever suggesting they weren’t, that they were just products of drift. Carl Zimmer seems to cited by many, many writers and bloggers as the best, most reputable, next best thing to a major working scientist, well nigh authoritative writer on evolution. I expect despite this acknowledgement of dissent, in practice the working assumption will be that Rinn/ENCODE are holding up their end of the controversy.
By the way I read this in Zimmer’s article: “Recent studies have revealed a wealth of new pieces of noncoding DNA that do seem to be as important to our survival as our more familiar genes. Many of them may encode molecules that help guide our development from a fertilized egg to a healthy adult, for example. If these pieces of noncoding DNA become damaged, we may suffer devastating consequences like brain damage or cancer, depending on what pieces are affected.” Sentence 1 says noncoding, sentence 2 says may encode, and sentence 3 say noncoding. Is this a typo or an equivocation or a switch in references from one sentence to the next I missed?
@2 stevenjohnson2
The confusion is in the use of the word coding and non-coding. Non coding DNA generally refers to DNA that does not encode for protein. However some of that DNA codes for RNAs that themselves serve to regulate the expression of proteins.
I suppose Carl Zimmer will now demand that you produce incontrovertible evidence that they don’t exist.
Checkmate, so much for scientism.
lorn, your comment makes no sense to me; could you try rephrasing it?
Non-coding also includes genes encode for protiens, but aren’t accessed during the lifetime of the individual. That is Junk DNA. False positives for ENCODE.
Slowly read the post, and then my comment. I’m sure you will get it. It isn’t complicated.
Explaining it would ruin the effect.
Oh, you mean your babbling? Explain yourself, as only you know what you said and meant.
I people knew the % of body weight that is genomic DNA, they would think it a big deal. I could triple my DNA and not notice it.
lorn@7, nope, even after rereading, your comment makes no sense to me.
@6 Nerd
Non-coding DNA can include regions of a gene that don’t encode for protein (like 5′ and 3′ UTRs and introns) but you would be hard pressed to find a gene encoding a protein that wasn’t expressed at some point during the lifetime of the individual.
@4 lorn
TOUCHÉ
Nerd@6: I’m not sure if that meets the definition of junk DNA. To me the key to junk DNA is non-coding, not lack of expression in an individual — using a gene-expression definition, you end up with the conclusion that HOX-2 genes are functional in B-lymphocytes and junk DNA in T-lymphocytes, you also have to conclude that a female mammal’s randomly inactivated X-chromosome means all the genes in the X-chromosome are junk DNA in half her cells.
DON KANE IS HEXAPLOID!
The problem is that there are unused portions of the genome decaying slowly that theoretically could code for a protein, but don’t ever get used except by mutations. Groups like ENCODE call all of them functional regions. If the gene is never accessed, it isn’t functional. I simply use the individual as an example to define a false positive.
In reality, since populations are involved, it has to be more than a single individual that doesn’t access that gene for it to be considered non-functional, hence non-encoding. But if that gene isn’t expressed in 1000/1000 individuals, it can be considered junk DNA.
@14 Nerd
If a gene is expressed in <1/1000 individuals but its expression is associated with a disease, would you consider it 'junk' DNA? An evolutionary biologist might, I suppose, but I doubt a medical doctor would.
wut
You don’t know what you’re talking about. At all.
ForFuckSake Sven
you can do better
By definition, every active parasitic jumping gene MUST get transcribed, and bind with the necessary transcription proteins. These would also ping as positive on ENCODE.
So would every single damaged/decayed copy of a parasitic jumping gene that still retained enough of its original sequence to have some degree of affinity for the transcription apparatus. That could include a significant fraction of the LINES and SINES, all of which are incontrovertibly junk, and all of which would ping as positive on ENCODE.
Almost every base in the genome is at some point transcribed and almost all of these transcripts can potentially encode proteins. The latter merely requires an initiation codon and many of these are indeed occasionally translated into peptides. Both transcription and translation are relatively sloppy processes and produce a lot of material that seems to have little purpose except being degraded. This debate is not about expression or encoding for proteins. The debate is about which of these transcripts and peptides have a function. And more broadly: what percentage of the genome is functional (including regulatory elements).
The difficulty is in the term function. There are many phenomena that have a measurable effect, but not what most scientist would consider a function. For instance if a transcript has a marginal effect on the expression of other genes but produces no noticeable effect on the cell or organism, is it then functional? By that definition everything is functional since merely transcribing random DNA will necessarily titrate some of the transcription machinery away from other genes. Hence, this definition of function is stupid. ENCODE unfortunately chose to define function as ‘biochemical activity’ which is very close to this definition. This has lead to a lot of confusion, because this is not what most people associate with the word ‘function’.
The nuances of ‘function’ is a too much to get into here, but one should recognize that function is to some extent a gradient. This goes from undisputably functional genes (i.e. any mistake in this gene is always lethal) through transcripts that only produce measurable effects under very specific conditions or have effects that are measurable, but neutral all the way to ‘biochemically active’ regions.
Anyway, with some exceptions, most people in the field believe that there are tons of non-functional (or ‘junk’) material in our genomes, but that there are still undiscovered functional non-coding transcripts, short peptides and regulatory elements. The controversy lies in how much remains to be found.
@18 Amphiox
90+ known human disease causing retrotransposon insertions as of 2014. They may be evolutionary junk but they aren’t medically irrelevant.
No one is claiming that no junk DNA is ever medically relevant.
Indeed, no one is claiming that junk DNA is never EVOLUTIONARILY relevant, since there are known examples of de novo protein encoding genes as well as regulatory elements evolving from junk DNA sequences. Icefish antifreeze proteins being just one example.
But that is a topic for a different discussion, one that has ZERO relevance to the idiotic definition of “functional” used by ENCODE.
@21 Amphiox
I read a paper a few months ago in which investigators reported that by using binding of two proteins (p300 and RNApolII) along with a half dozen different histone post-translational modifications and looking for the overlap of protein binding with some histone medications ( and exclusion of others) they could identify enhancer elements in sea anemone with at least 70% accuracy. Pretty much in agreement with what has been shown in mice, humans, worms and flies (as established by their respective ENCODE projects). So what exactly makes the definition of functional used by ENCODE idiotic?
Darn that spell-check. Histone modifications not medications.
@chris61
I fail to see how the accuracy of enhancer identification is relevant to the general definition of ‘function’. However, if you read my previous post it clearly states the problem with the ENCODE definition of ‘functional’: it does not align well with what most people associate with the word ‘function’.
It uses non-specific proxies for defining functionality. Not function itself, a NON-SPECIFIC PROXY. Any scientist calls that stupid. What is your excuse for using an idiosyncratic defintion? There is none.
@24 eviledv
The point is that they have identified biochemical signatures of regulatory elements that appear to be common across the animal kingdom. DNA sequence alone doesn’t identify a regulatory element but these signatures do much of the time. It’s an opportunity to study an aspect of evolution (i.e. the evolution of various classes of regulatory elements) that wasn’t available before. Whether or not most people associate those biochemical signatures with the word ‘function’ is really beside the point.
@25 Nerd
In fact Nerd, if you check the literature you’ll see that many, many scientists are too buy making use of the data generated by ENCODE and mouse ENCODE and modENCODE to pay any attention to a discussion of whether biochemical signatures constitute ‘functionality’. Whether or not you call them functional, they are proving to be extremely useful.
@chris61
The question is what extent of the genome is non-functional/junk DNA. If you think a biochemical signature is proof of function you are mistaken. Also, as a side note, ENCODE did not (and never claimed) they invented evolutionary studies of epigenetics. Leaving aside the identification of enhancers, which is in itself a semi-controversial field, the vast majority of ‘functional’ elements in the ENCODE study are identified by the fact that they are transcribed. This includes introns and transcripts that are likely byproducts of a sloppy mechanism. This is not a very helpful definition of ‘function’. Hence:
1. A large part of the genome is likely DNA with no function as understood by most scientists
2. ENCODE muddled this debate and confused the public by redefining ‘function’ as anything that is biochemically active
As for the usage of ENCODE data. Of course scientists are using their data. I am using it too. It’s a great resource. However, many scientist were not thrilled by the way the study was framed.
@27 eviledv
I think a biochemical signature is indication of biochemical activity. How that biochemical activity contributes to the phenotype of the cell or organism is a matter for experimentation.
Of course ENCODE and its off-shoots didn’t invent evolutionary studies of epigenetics but they have expanded them in a big way.
How can one leave aside the identification of enhancers when changes in gene expression, which are largely driven by enhancers, are probably the largest source of variability between closely related species?
Surely you aren’t suggesting that introns aren’t functional?
You talk about transcripts that “are likely byproducts of a sloppy mechanism” but the point is that having identified these transcripts scientists can experimentally test whether they are functional or not.
The idea behind ENCODE was to look for certain general patterns that were associated with various known functional chunks of DNA, where functional means encoding, promoting or otherwise doing something that we already know something about. In some ways this is a good idea, because the odds of something functional looking like something else functional should be better than chance. The big weakness is that it captures every face drawn on a planetary surface, even on planets where nobody ever drew faces.
Given how frequently evolutionary biologists find new genes developed from redundant copies of older ones, one would expect there to be all sorts of redundant copies in various states of repair, sort of like cars in a poor, but handy person’s driveway. A few might actually be drivable, others might just need a few parts, most might be real junk, but a casual observer, like an ENCODE statistical test or a county junk car enforcer, would recognize each of them as an automobile.
I don’t get the resistance to the idea of junk DNA. The mechanics of evolution almost demand it. While there may be costs to copying extra genes, there are also risks to cutting sequences out too aggressively.
P.S. Why can’t I ever think about junk DNA without remembering Saturday Night Live’s old metric alphabet, the decibet, which condensed all of the so called “trash letters”, P through Z, into a single “easy recognizable dark character”? (http://snltranscripts.jt.org/75/75rdecabet.phtml)
@chris61
1. Biochemical activity in general is a meaningless definition of function. The whole genome is biochemically active. Hence, claiming anything about junk DNA based on this general definition is also meaningless.
2. I’m leaving aside enhancers because they are largely irrelevant for this discussion. The main part of “functional” regions are labeled so because they are transcribed. I’ll be happy to discuss enhancers, but that’s a separate discussion.
3. Of course I’m not suggesting that introns are non-functional. I didn’t think it was necessary to spell this out but the large parts of many/most introns can be deleted with no consequence.
4. To identify a transcript is something entirely different than claiming that it is functional.
This is getting a bit tedious. You are trying to defend the ENCODE definition of “function” so please explain how bichemical activity is a good substitute for functional.
@31 eviledv
What do you consider a meaningful definition of function and will your definition help me identify non coding DNA sequences that are important for regulating gene expression?
Also when you say large parts of many introns can be deleted with no consequence, what proportion of introns are you referring to, how do you propose to identify those introns that can’t be deleted without consequence, and what do you mean by “no consequence” anyway?
Why do you need to call it “function” at all?
Why misappropriate a term that already has an established and accepted definition, knowing full well that in doing so you confuse the issue and give ammunition to anti-science types?
Invent your own goddamn term and give it exactly the definition you want to give it, to use for whatever purpose you wish to use it for.
Why do you misappropriate perfectly defined words and apply an idiosyncratic definition that no other group of scientists use, unless you are either deliberately lying, or you are trying to sow confusion for political purposes?
Define your own words, but leave those previously defined and understood as they are.
ENCODE has bad reputation just for such arrogant behavior. Get ENCODE to stop that behavior, instead of defending it.
@33 Amphiox
So what is this established and accepted definition?
@33 Amphiox
Also, if you’re concerned about giving ammunition to anti-science types, perhaps you should have a word with Richard Lewontin quoted in on of PZ’s recent posts as saying ” I think the admission of necessary ignorance of historically remote things is the first rule of intellectual honesty in evolution.” That is a quote just waiting to be willfully misinterpreted.
@32 chris61
You seem insistent on derailing this conversation and focusing on other questions than this discussion’s topic. I believe I have given you reasons why defining ‘function’ as ‘biochemical activity’ is 1) not helpful and 2) does not say much about junk DNA. The main argument being that this labels the whole genome as functional if you just sample/sequence deep enough. This is valid regardless of enhancer identification or what proportion of a given intron can be deleted.
The simple truth is that transcription, epigenetic marks and binding sites are not in and of themselves proof of any meaningful biological activity. For instance believing that all lncRNAs are functional (which you seem to imply with your views) is a rather extreme position and not in line with what most people in the field believes. Sure, these signatures can indicate function and if one is trying to identify functional elements one should prioritize regions that have one or more of these. But I reiterate that observing a transcript is not evidence of that transcript being functional. I strongly suggest you read Sean Eddy’s piece in Current Biology which gives one of the more insightful and measured comments on this.
I think I will sign out of this discussion now as it is leading nowhere.
@37 eviledv
Thanks for the conversation although I’m disappointed that you wouldn’t clarify what you consider a meaningful definition of functional.
The definition everybody else uses. The gene is turned on at some point in the life of that organism. If a gene is never turned on, it is non-functional. Given the known mathematics, genes are considered non-functional until shown functional with solid scientific evidence. Which excludes the use of proxy’s to define activity/function.
@39 Nerd
Turning on a gene means to transcribe it and most, if not all, of the genome of humans (along with flies, worms mice etc.) is transcribed at some point in the life of that organism. So you appear to be telling me that ENCODE’s definition is the definition that everybody else uses. In that case, why is everybody else complaining about ENCODE’s definition?