I understand that Los Angeles is a barren wasteland of ennui in February, unlike Morris, Minnesota, so in a fit of altruism I’m bringing a spark of Minnesota excitement to boring California. I’ll be speaking at CFI LA, in a talk that is sure to piss off a lot of people.
Bad Biology: How Adaptationist Thinking Corrupts Science
11 a.m., Sun., Feb. 15 – Hollywood; 4:30 p.m. – Costa Mesa
In 1979, Gould and Lewontin published an important paper, “The Spandrels of San Marco and the Panglossian Paradigm: A Critique of the Adaptationist Programme”, in which they deplored the narrow focus on seeing evolution as only a process of adaptation. They further criticized the field of evolutionary biology for perpetuating this pattern of flawed thinking.
Biologist PZ Myers bemoans that their warnings have gone unheeded in some biological sub-disciplines. He’ll be discussing a few examples of bad evolutionary biology, including the “human biodiversity” movement that is little more than a collection of pseudoscientific rationalizations for racism; evolutionary psychology, which seeks to explain modern human biology with imaginary scenarios of adaptive constraints from 10,000 years ago; and the ENCODE project, which was an obvious and overt exercise in adaptive bias applied to genomic data. A common thread among these examples is an excessive reliance on adaptationist thinking and a lack of appreciation of the diversity of mechanisms underlying our evolutionary history.
Although, to be honest, another factor is that February tends to be our very coldest, bleakest month, so a weekend in Southern California seemed like a good idea.
It’s too far for me, but have a good time! Safe travels.
I’m further north in Silicon Valley: the number of visitors we get from the various MN outposts of high-tech companies peaks around early February each year.
Ooh PZ, some folks will be very angry that their justifications for their bad behavior is soundly rebutted.
I’d be more impressed if PZ were going to Winnipeg or Novosibirsk in February. SoCal is a snowbird migration stop in February.
I’ve done similar excursions from cold N.E. to Sunny S.F. It’s fun to get out of the cold for a week, but then going back to the cold (from the warm), makes the cold feel even more frigid. Good luck PZ, the weather shock puts jetlag to shame.
I might actually be able to make it to that.
re @2:
[raising hand]: My guess is Seagate is a large part of that. HQ in Minneapolis suburb (Shakopee) with a recently acquired office complex extension (via Maxtor/Quantum) in Milpitas, CA. Feb is perfect opportunity for Shakopeeans to hit Milpitas (for the warmth).
Make sure you enjoy the rush hour traffic!! It can’t be beat!!
Here, I’ll burn one last bridge.
Please do some homework first. Please be able to support your assertions.
Please don’t pretend to be giving a science talk when it’s obviously a political rhetoric talk.
Nobody considers “the human biodiversity movement” to be a “biological sub-discipline”. Nobody.
As for evolutionary psychology, the well is already poisoned by the “imaginary scenarios” crack. No piont in expecting an objective analysis there.
ENCODE had little ot nothing to do with adaptationism; the whole kerfuffle was due to the use of a ridiculously broad (and non-evolutionary) definition of the word “functional”.
Waving around a copy of Gould and Lewontin is not science. It impresses no scientist other than Larry Moran. Science is the logical interpretation of data. Will there be any data presented?
If your goal is to fight the perversion of science for political/rhetorical purposes, then it looks a lot like you are about to score an own-goal.
But have fun preaching to the choir.
Happy Newt Year.
Tedious. That’s how I’d describe Chas. Not trolling, not irritating or even mildly annoying anymore. Just a persistent, predictable, unnecessary, snarky little speed-hump in every thread. So. Fucking. Boring.
You’re lucky it is on Sunday, otherwise you’d never make it from Hollywood to Costa Mesa (assuming a 3 hour talk). :P
Also, nothing cool ever happens in Costa Mesa, so no loss if you skip that part of the trip and stay in Hollywood (beebop over to West Hollywood for the real action!)
I might just be able to stop by on that date PZ. And Chas, thanks for stopping by and pissing on everyone’s cornflakes.
Oh, at first I thought that was the point. I’m not even from Minnesota, but winters in coastal California are really boring. I’m looking out at a sunny backyard in the SF Bay Area right now. It is a bit cold, though. It may have (gasp!) gone below 50F last night.
Maybe what LA needs right now is an indoor ice-fishing complex. There’s got to be something better to do in February than walk around looking beautiful, right?
Well Chas, other folks has made the science, such as it is, political. Therefore, politics must be discussed during the debunking. So, why are you so adamant about sticking to the science, and not address the political problems caused by people misusing the science? It just doesn’t make sense to me.
@7: twas brillig: got it in one – the HDD industry has many locations in common, in the US: Silicon Valley, Longmont in CO, in and around Minneapolis, and not so much now, Massachusetts.
I’ve done the reverse winter trip to MN several times. Even though Seagate purchased Maxtor (including the old Quantum), they don’t use any of the buildings in Milpitas from that acquisition. They are in the process of consolidating into the (completed just before they went bust) Solyndra manufacturing building in Fremont as well as having moved their HQ from Scotts Valley to Cupertino (to be closer to Apple and other customers).
I changed companies a few years back and now quite regularly end up in Japan, which should give you a good idea of my current employer. Overseas, Singapore, Shenzhen in China and Thailand are the obvious other common areas. The Thai location was made infamous in the industry with the flooding a few years ago which submerged something like 50% of the industry’s manufacturing capacity for making recording heads and greatly spiking prices. For a fun (in hindsight) perspective on the implications of the flood see here: https://www.backblaze.com/blog/backblaze_drive_farming-2/
It might be clear and cool here then, might be rainy (yay rain please), but it will definitely not be snowy.
I hope you’ll address Dan Dennett’s objections to Gould and Lewontin’s paper that were made in his book Darwin’s Dangerous Idea back in 1995. It turns out that spandrels (or pendentives) are also adaptations to an architectural problem of supporting arches and domes, and there were precedents to them that were tried and found wanting. I think the real problem here isn’t adaptation per se but greedy reductionism that seeks to claim explanations for behaviors that have a far more complex cause.
@14 Nerd (to Chas):
Good question. The CFI is, after all, a political organisation:
So, if the goal is to literally change society, how would it be at all out of place for PZ to discuss the wider, non-scientific problems that can stem from bad or flawed or misread science? It seems that that sort of discussion would be perfectly appropriate at a CFI event. In fact, just sticking to the science might only be telling half the story.
Regardless, how dare someone roll in and first prejudge, then presume to prescribe the content of a speaker’s talk? White-hot fricking arrogance in action.
But also quite tedious.
Chas:
It is dishonest to pretend that the adaptationist model has no political implications. Both sides do. Yet somehow it’s only when views contrary to the status quo are presented that suddenly we get these pointless accusations of “politicizing”.
They do. James Watson does. They tend to put themselves on the pedestal of being the only pure biologists, unbiased by political correctness. They sound quite a bit like you, actually.
I’ll be contrasting evo psych with good human evolutionary biology. I’m afraid evo psych specializes in the imaginary.
Why did they assume that every bit of the genome was functional? It wasn’t because they had good data supporting that conclusion. It was because of the default assumption that every feature was there because natural selection put it there.
It was an ideological conclusion. I know you think only one side is ideologically driven, but you’re going to have to broaden your horizons a bit.
Nah, I’m just going to wave around a copy of Gould & Lewontin, pound the lectern, and declare all who disagree to be enemies of the state who must be shipped out to a gulag.
Say, weren’t you just sneering at your perception of a well poisoning? What do you call this?
I think that Gould/Lewontin’s conclusion, that we ought to provide far greater consideration of evolutionary mechanisms other than simply natural selection, is better science than the continual denial of the importance of chance, or the assumption of selection as the null hypothesis.
I don’t anticipate a purely laudatory audience. Quite the contrary.
I live beyond 66° north, so your cold and bleak February sounds just fine. ;)
As for the topic, though: It always intrigues me somewhat that people who admit science is in part a political endeavour are accused of ‘politicising’ science, while people who pretend science could replace politics with “rational policies” or the like, get away with it. (Except for being criticised by people who “politicise science”.)
It always looks uncannily like the relationship between religion and politics to me.
@PZ
They didn’t assume anything – they showed experimentally that the majority of the genome is functional according to their definition of function (i.e. having some associated biochemical activity). Whether or not that biological activity contributes to phenotype is testable and is being tested.
And where did that entirely useless definition come from? Why did they use it?
What makes it an entirely useless definition? Entirely useless to whom? Certainly not to those attempting to annotate the human (and other) genome(s) in efforts to identify regulatory elements.
Try reading Eugene V. Koonin, The Logic of Chance, Then Nature and Origin of Biological Evolution. The discussion at where I am at the moment is why multi-cellular organisms have genomes larger than expected. Why? The energy costs of cleaning up the genome, is larger than what is gained in streamling the genome, as non-coding areas might be reused in the future. And it often is.
It would be expected that some areas of the genome, even most of it, might not be active at any time for an idividual.
@24 Nerd
In general, when a properly controlled and validated experimental observation fails to agree with a theoretical expectation, further experimentation is called for and should that experimentation still fails to agree with a theoretical expectation, theorists are expected to modify their theories to accommodate the experimental data, not vice versa.
Read Dan Graur’s thorough excoriation of the ENCODE project. An effort to catalog functional elements of the genome that has such a loose definition of “function” that on the first pass they could say 80% does something (although they couldn’t say what, other than that a protein sticks to it sometimes) actually impedes efforts to identify function. It’s the false positive problem…and hoo boy, did ENCODE embrace their false positives.
How do you know which ENCODE results are false positives and which are real? I read Dan Graur’s immortality-of-television-sets. It’s amusing but doesn’t as far as I can see really address any of the questions that the ENCODE project looks to address.
And I haven’t seen this real data. Just the claim that the entire genome active. So, the theorists don’t have to change anything at the moment.
Easy, be scientific, and presume they are false until proven real.
@28 Nerd
If you haven’t seen the ENCODE data, know what it consists of or how it is acquired, I would suggest that you aren’t in a position to know whether it is false or to decide the grounds on which it should be verified.
I’ve been a scientist for 40+ years, and read a lot of science news. Take your bullshit and shove it where the sun don’t shine.
There is nothing wrong scientifically with acknowledging a data from a poor test method is likely false positives, and further testing is need to confirm if it is a real or false positive. That is my point. It is false until proven real. But certain folks get a burr under their saddles if their preconceptions, and the entire genome being active is an obvious presupposition, isn’t treated as gospel.
@30 Nerd
So what, in your opinion, makes it a poor test method?
And I don’t think you’ll get one. Ed Clint lives in L.A., and I’ve seen him at CFI. And we have our own contingent of Status Quo Warriors.
False positives, as has been mentioned. Why don’t you folks doing the work see it as a problem? Everybody else does.
I’m a lurker, but will try to make it since it sounds interesting.
Also interesting, if you have free time in LA, is visiting the Watts Towers, which will prove to you once and forever that Los Angeles is not any kind of barren wasteland but rather a place of great creativity.
@33 Nerd
False positives were indeed mentioned previously although no specific examples were provided. Would you care to provide some?
PZ mentioned some. but google showed pages of data with articles like this:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3431496/
So, will you quit pretending your data is pristine and no false positives occur?
I’m not a biologist so I’m possibly missing something obvious here, but …
Using ENCODE’s definition of function wouldn’t have we known that almost 100% of the genome is functional before even starting the ENCODE research?
By that I mean we know that except for the end of the telomeres, almost all of the DNA gets copied during mitosis, which I would think qualifies as a biochemical activity and thus as functional according to ENCODE, thus leading to a near 100% functional genome.
If replication during mitosis is somehow excluded* from biochemical activity, then what reasoning was used to do so and are LINE and SINE replication excluded too?
* which I expect it is as ENCODE was claiming about 80% of the genome was functional when it first released its results.
@36 Nerd
So the link you provided links to a paper establishing objective and reproducible criteria for assessing the quality of ENCODE data. That is hardly the feature of a poor test method. In any case I note that you still haven’t provided a specific example of an established false positive.
@37 Julien
The feature of DNA replication that has functional implications of interest to ENCODE is its timing during the cell cycle. Active genes generally replicate earlier than inactive genes.
To do in LA or a few hours away:
La Brea Tarpits museum.
Sequoia Nat. Park (probably cold and snowy in December, but hey, you’re from MN…)
Wow. I had to look for myself.
On the Immortality of Television Sets: “Function” in the Human Genome According to the Evolution-Free Gospel of ENCODE
So you are “churchgoer” if you have ever been in a church even once. Or humans are “air movers” because we move against air. Or your furniture are “carpet compressors”.
How about we say something is functional when you actually identify the function beyond the fact that it sticks to some shit? As far as I’m concerned no one can say something is functional until they have actually identified what unique role it has in the cell (and “sticking to that antibody that one time” does not count).
It’s nice to see that Chas is still making the same dumb mistakes. Is it still projection when someone does the same damn thing they are accusing the other person of doing mere sentences earlier or later? It has to at lest be a little bit hidden right?
So you’ve seen the talk then? Oh right, that’s your assertion that the talk yet to be given is a “political rhetoric talk”.
EXACTLY.
You don’t seem to realize it, but that’s the problem with ENCODE.
Whoever thinks that “science” is unpolitical knows so little about the history of science that they shouldn’t be allowed to use that word.
remember that there are four seasons in LA – fire, flood, earthquake, and riot. feb-march is the rainy season.
costa mesa is far past the orange curtain. please make sure you know how you’re getting there from hollywood. yes, it’s a sunday, but it’s still a minimum of 1 hour.
Being from Southern California, I can assure you that Los Angeles is a barren wasteland, built in a desert, every month of the year. To get from San Diego to San Francisco, everyone tries to plan going through it at 2 or 3 in the morning.
@41 PZ
You seem quite convinced that there are false positives in the ENCODE data but if you can’t identify them, how do you know that is the case? Maybe 80% or more of the genome is functional (i.e contributes to phenotype). On what basis would you assert it isn’t? Moreover, from a purely practical point of view, annotation of genomes with biochemical signatures is serving quite nicely to identify potential regulatory regions that can be experimentally tested for their functionality. The annotation has predictive value and that itself is a useful function.
On what basis, other than presupposition, do you assert the entire genome is active? Look at yourself cricket, you are the one making the mistakes, and ignoring the data.
Chris, my paper labeled certain data as false positives. You can’t be wrong. The first lesson of being a scientist is you are the easiest person to fool, and must take that into account. You have fooled yourself.
Been there, done that. And it failed. 4:30 AM, the I5 packed full with cars not moving, because someone’s pickup truck burst into flames.
@46 & 47 Nerd
The assertion was not for the entire genome, just for 80% or more of it and was based on experimental data. Data is not presupposition.
Your paper pointed out criteria for identifying likely false positives. It didn’t (because it couldn’t) demonstrate that any of those data points represent an actual false positive (or a true positive for that matter). Your paper did not address what percentage of the human genome is active or functional.
-February is our month of deluge.
-An experience you will never have anywhere else: The Museum of Jurassic Technology (culver city). To make the most of your visit, know as little as possible – its secrets are best … discovered.
-sans deluge, our Getty museum(s) have brilliant architecture and view-ness.
-if your timing is right, there is a fantastic art-walk event each month in the dowtown area. Great restaurants and open galleries!
-LACMA is big enough to have something interesting, somewhere. Tar pit natural history museum is adjacent and has unique angle: “these died horribly … Here”.
-A space shuttle resides at USC and you can reach it via shinynew metro. Science museum there often has touring shows of quality as well, though otherwise more broad than deep.
-if you can make it down to long beach the aquarium of the pacific there is quite excellent. Region surrounding has happyfun harbor / restaurants. Careful timing needed to avoid rush hour vexation ( but can use metro rail).
-if the troubadour theater company is playing during your stay, see them if you can!
@chris61
Ass-backwards. Stop trying to shift the burden of proof.
They assert that “sticks to something” = “functional”. I choose to disbelieve that until they provide something for functionality that includes working with the individual proteins (or RNAs or…) and providing specific evidence for what each does. The same way they do for the individual players in photosynthesis for example. At best I would call this “first pass putatively functional elements”.
The false positive issue is secondary (it’s ENCODE’s obligation) and false positive are a very common problem with any sort of antibody based purification assay. Simply [googling antibody and “false positive”] tosses up tons of examples that are even in currently approved products. Some things like ribosomes are notoriously sticky, and there are even proteins like ZDS1 and 2 that stand for “Zillions of Different Screens”. I pulled those two out of a screen myself.
@chris61
Data that amounts to “It sticks to something” in conditions that are imperfectly approaching physiological at best, and independent of many things that may normally influence the binding but now washed away. We have no idea how many of these things are only sticky because of the conditions and it ENCODE’s job to determine that. If Nerd of Redhead’s paper shows methods of identifying false positives then ENCODE has work that they better get to.
That’s not what you said above, where ALL the genome is active. In case you aren’t terminally arrogant, you need to consider those who are questioning the data, and why they question it.
From the Wiki article on ENCODE:
[26] Graur D, Zheng Y, Price N, Azevedo RB, Zufall RA, Elhaik E (2013). “On the immortality of television sets: “function” in the human genome according to the evolution-free gospel of ENCODE”. Genome Biol Evol 5 (3): 578–90. doi:10.1093/gbe/evt028. PMC 3622293. PMID 23431001.
[27] Moran LA (2013-03-15). “Sandwalk: On the Meaning of the Word “Function””. Sandwalk.
[28] Gregory TR (2013-04-11). “Critiques of ENCODE in peer-reviewed journals. « Genomicron”. Genomicron.
So you can see there is considerable criticism of the methods, and their meanings. To claim that every hit is an active gene is simply ludicrous due to false positives. Get some humility, and realize dogma of the project when you see it. Skeptics fight such dogma.
@51 Brony
Exactly! It’s a screening test and it identifies putatively functional elements. You don’t dismiss screening tests as being useless until you’ve done the tests to see how useful they are. The tests are being done based on ENCODE data and the tests suggest that ENCODE data has predictive value in identifying regions of DNA with regulatory function. Of course there will be false positives and false negatives as well because that is the nature of biochemical tests. Nobody is suggesting throwing out DNA sequence conservation across distantly related species as a means of identifying functionally important sequences even though it has false positives and false negatives as well.
A screen that reports nearly 100% positive results is not a screen. Imagine a cancer diagnosis tool that announced that everyone had cancer. How useful would it be?
Do you even understand the concept of a false positive? Do you work with the ENCODE project?
@55 PZ
The screening aspect of it isn’t in which DNA sequences have ANY biochemical signature but in identifying sets of biochemical signatures that correlate with known regulatory regions (like promoters or enhancers) and then using those signature sets to identify DNA sequences that can then be subjected to further testing for promoter or enhancer activity.
@53 Nerd
In fact I’ve been very careful not to state that ALL the genome is active.
Yeah, you don’t understand what a false positive is.
Then it should be reported that 80% of the genome is potentially functional, not that it is functional, as you claim. And you haven’t been so careful as you think.
Example of something that ENCODE would find potential functional: ancient retroviruses, used to determine primate lineages, but most likely not functioning during the lifetime of the animal. All false positives.
@58 Nerd
http://www.pnas.org/content/111/17/6131.full.pdf+html
PZ – please post a link to the talk if/when it’s available online!
@chris61
Then if “putatively” does not come along with the word “functional” every single time they are committing scientific malpractice as far as I am concerned. That “80% functional” takes the word functional and associates it with almost nothing in the way of specificity outside of “sticks to some shit”. Not even remotely good enough in the human society we have that abuses science the way it does. That, “The NRXN1 locus is functional” and “80% of the genome is functional” are equivalently specific is an atrocity. This is a cultural issue as well as a scientific issue and if you can’t functionally address that you are not helping.
Ass-backwards again.
ENCODE’s results are to be taken as suggestive only if they can be correlated with other data. It’s not dismissal, it’s ignoring for good reason until something else gives a reason to look at the ENCODE data set. If I tried using an ENCODE result to propose a project I would be looked at very skeptically. Screens are only to be used with other data. Their claims of (80% functionality) do not get this reality across and are utterly misleading.
A claim of function for 80% of the genome without independent evidence of function from specific regions is ridiculous.
Additionally a claim of function for 80% of the genome without directly addressing the issue of false positives in that same claim is ridiculous. If one is talking about 80% of the genome I better see an exhaustive list of ALL of the ways false positives occur, and what they specifically doing to control for them. Without that ENCODE sits there unused until a lab has some other reason to go look at what they have.
It’s a good thing I never did that then.
This quote from your the paper you cited Chris is the heart of the issue. From Brony’s posts, we are both of the scientific theory that you must show each area you claim is active from the ENCODE tests is actually active in humans at any time during their lifetime to call it functional. There is an awful lot of duplication and degrading retrovirus material in the genome, and I would speculate a large percentage is considered non-functional, in that it doesn’t turn on and become active during the lifetime of a human. Which is the hard core and absolutely legitimate definition of being functional. You see, hard evidence to back up the claims is required.
@61 Brony
Have you read any of the ENCODE papers? There are in fact pretty exhaustive descriptions of what is being done to maximize signal to noise. But your request of a list of all the ways in which a false positive could occur is meaningless unless you first specify what you mean by a false positive and to do that you first will have to define the criteria by which you acknowledge a ‘true positive’. What do you mean by a true positive?
As far as the data sitting unused… well a quick search of PubMed or google scholar will demonstrate that isn’t happening anytime soon.
Simple, a positive signal in vivo, when there is no turn-on in vivo. Why are you so obtuse about such a simple concept?
He explained it, but you were too busy to acknowledge it. Conclusive evidence the gene it turned on during the lifetime of a human. Simple. Too simple for your mind evidently.
@ chris61
Context note: the issue centers around concerns (scientific and cultural) involving claims that 80% of the genome is functional. This is ENCODE’s claim and you are choosing to defend them. You should already be handing me a paper because if you are happy with how they have handled it you should be able to show me.
I will not go fishing around in your head. Especially since you had nothing to say about my comments on the specificity of the claim “80% of the genome is functional” compared to other common usages. It’s fucking insulting.
How about paper? If they have looked at false positives at 80% coverage you can expect just about every general type of false positive to occur. They have an obligation to list them, and show how they dealt with them.
Seeing as I mentioned how the ENCODE data could be used in my comment above this is meaningless. I implicitly acknowledged that it can be used with limitations on how it should be used. If you can’t actually address my comments accurately why should I trust that you know what you are talking about?
@62 Nerd
That is the point. It isn’t simple for a number of reasons mostly relating to the difficulties in defining what you mean by functional and then deciding on appropriate tests for functionality. Having done that, the tests themselves, unlike the annotation, can never be high throughput so they’ll take years to perform and will never be performed for the entire genome. In any case the bottom line is that the precise fraction of the biochemically annotated genome that should be regarded as functional, in the end, has no bearing on the utility of the ENCODE project.
What do you mean by active? Actively transcribed? Actually a lot of it is. Or do you mean active in some other way and if so, what way?
Then all claims that they are functional should be dropped. That is good science.
Chris being stupid:
in that it doesn’t turn on and become active during the lifetime of a human.
What part of being turned on did you miss?
In a computer program, if you have subroutine 3782, but your program never calls it, you have junk (non-fuctioning) code. It reads like it does something, but it doesn’t. Same for the genome.
You don’t seem to understand biochemisty and genetics, but yet you talk about it pretending to be an expert.
A little history is sometimes useful, IMO. Ernst Mayr wrote a paper back in 1983 replying to Gould and Lewontin that’s worth mentioning as part of the debate that took place back then, so here it is:
How To Carry Out the Adaptationist Program
Mayr mentions how after other algorithmic mechanisms like saltation (“hopeful monsters”), inheritance of acquired characteristics (Lamarckism), and orthogensis (the notion that there’s some sort of direction to evolution), what is left is chance and natural selection to explain the apparent design seen in nature. Obviously, natural selection plays some role (consider why the fur on polar bears is white, for one example) in this, and in fact, it would be foolish to deny it. What Gould and Lewontin were driven to respond to back in 1979 wasn’t adaptation as much as it was E.O. Wilson’s Sociobiology (1975), a text that for the most part dealt with insects, but aroused much ire when it got into speculations about human nature being similarly shaped by evolution. Keep in mind, this was the 1970s, and there was a lot of drama that resulted that still reverberates to this day.
@ 65
I have told you and paper specifies how they define functionality and how they go about maximizing experimental reproducibility and signal to noise. Neither they nor anyone else is ever going to ‘prove’ what percentage of the human genome is or is not functional in the way you seem to mean it. The problem lies in defining functional and what it means for non-coding sequence to be functional. You speak of common usages but that can mean a lot of things and I don’t know exactly what you mean by it. If you’re asking, can it be deleted without an obvious effect on phenotype? Some of it can and some of it can’t but most of it hasn’t been tested because those aren’t easy or cheap tests to perform. But even if it has been tested and it has no obvious effect one still has to ask, how carefully was the phenotype analyzed. There are human polymorphisms that appear equally functional until something like HIV comes along and what do you know, one allele makes a person relatively resistant to viral infection while another doesn’t. Is that a functional difference? Only if you look in the context of a human population exposed to HIV. One also has to ask whether genetic redundancy might be coming into play in identifying functional non coding sequences. Among coding sequences where potential redundancy can be easier to recognize (based on sequence homology) there are plenty of gene families (particularly among transcription factors) in which deleting one family member has little effect on phenotype but deleting two or more can have a profound effect. Nobody would argue that the individual genes aren’t functional. Non-coding sequences probably have functional redundancy as well but if they don’t have sequence homology, it’s not going to be easy to identify functionally redundant elements.
@67 Nerd
Actually Nerd you don’t seem to understand genetics or rather, you seem to think that the human genome is perfectly analogous to a computer program. It isn’t. DNA is not ‘read’ in all the same way. We know (mostly) how to read coding sequences. They are transcribed to RNA and translated to protein and we can mostly detect when that happens. There are technical limitations to detection so it is at least theoretically possible that a gene would be expressed transiently during embryogenesis in such a restricted manner that it wouldn’t be detected. We don’t know how to determine when non coding sequence is active. If by activity of a regulatory sequence you mean when it is regulating something, that’s all well and good but how do we recognize when a non coding sequence is regulating something. That’s the point of annotating the genome, to determine how to use biochemical signatures to identify when a regulatory sequence is active.
@chris61
So you don’t know what the people you are defending do with false positives. I have it open in front of me and you do not appear to know the contents given your comments (hint, it’s September 2012). Additionally I’m looking at an analysis of the quality of these data sets. You are defending them. After this comment I will have nothing more to say to you if you can’t start getting specific about what you know of false positives and encode.
You have also not told me how they define functionality. Please point to the comment where you tell me such.I’m looking at it right now.
Given your inability to recognize that I acknowledged that ENCODE data can used above I’m not going to trust your ability assess what I mean when I say functionality, or your familiarity with ENCODE.
See my NRXN1 comment that you utterly ignored. We can look up functional details about that protein. The word “functional” in the context that ENCODE uses it has little resemblance beyond that it stuck to some things.
Exactly. You need all of those details to figure out what functional means in the only use that can actually be useful. What the piece of DNA specifically does in the overall context of metabolic processes over time. I can add enzymatic function assays, experiments to look for and confirm regulatory regions, experiments to look for binding partners, and much more. Even one of these is better than what they are offering.
Hearing you complain about how hard it is pretty revealing. Functional means nothing else beyond sticking to something here.
Same as above here. This is what you need to really know that something is functional. Otherwise all you can say is that X has been found to bind to Y.
To be more specific in 71 when I say.
<blockquote cite=""You have also not told me how they define functionality. Please point to the comment where you tell me such.I’m looking at it right now.</blockquote cite=""
"it" refers to information on definitions of function from ENCODE.
Nope, won’t argue with that, but my point if the piece of the genome you call “fuctional” is never used during a person’s lifetime, is it really a functioning human gene? I don’t think so, nor do ENCODES critics. And this the point you are obtuse to the point of stupidity about.
Thats not what you say. You say a gene is functional based upon in-silico and in-vitro data, which can be misleading due to false positives. The gene is never activated during a person’s lifetime. It can be non-fuctional in-vivo, where is the only place it counts. Your artificial definiton of functional isn’t a functional as you pretend it is.
What ENCODE needs to do, is to change their language and definitions to very conservative science. Either show in-vivo, or don’t claim functionality.
@71 Brony
ENCODE defined functional as DNA sequence reproducibly bound by some specific proteins or transcribed or the histone proteins within the nucleosomes binding in and around the sequence carry specific post translational modifications or the DNA is not bound by nucleosomes at that particular site; all of that occurring in some specific cell lines grown under specified conditions. There are actually a number of biochemical signatures annotated by ENCODE and unlike other tests of function, they are amenable to high throughput analysis.
That was your NRXN1 comment. If you really want a comment on it then mine would be this: What ENCODE means by functional is defined in their papers and it isn’t equivalent to saying the NRXN1 locus is functional (which is defined according to a whole different set of criteria and tests).
@73 Nerd
We are speaking different languages. The sequence is functional as defined by specific biochemical signatures specified in multiple ENCODE papers. Is the sequence ‘used’? Again what do you mean by used? Used for what? Transcribed into an RNA? Probably but not necessarily and there is no evidence that transcription is a requirement for function. Is it translated into protein? Probably not as less than 1% of the human genome is translated but again translation is not a requirement for function.
Here’s an example for you of a functional non-coding sequence identified using the techniques of ENCODE. Note that epigenomic annotation is basically what ENCODE does.
Recessive mutations in a distal PTF1A enhancer cause isolated pancreatic agenesis
Michael N Weedon, Inês Cebola, Ann-Marie Patch, Sarah E Flanagan, Elisa De Franco, Richard Caswell, Santiago A Rodríguez-Seguí, Charles Shaw-Smith, Candy H-H Cho, Hana Lango Allen, Jayne A L Houghton, Christian L Roth, Rongrong Chen, Khalid Hussain, Phil Marsh, Ludovic Vallier, Anna Murray, International Pancreatic Agenesis Consortium, Sian Ellard, Jorge Ferrer & Andrew T Hattersley
Nature Genetics 46, 61–64 (2014) doi:10.1038/ng.2826
The contribution of cis-regulatory mutations to human disease remains poorly understood. Whole-genome sequencing can identify all noncoding variants, yet the discrimination of causal regulatory mutations represents a formidable challenge. We used epigenomic annotation in human embryonic stem cell (hESC)-derived pancreatic progenitor cells to guide the interpretation of whole-genome sequences from individuals with isolated pancreatic agenesis. This analysis uncovered six different recessive mutations in a previously uncharacterized ~400-bp sequence located 25 kb downstream of PTF1A (encoding pancreas-specific transcription factor 1a) in ten families with pancreatic agenesis. We show that this region acts as a developmental enhancer of PTF1A and that the mutations abolish enhancer activity. These mutations are the most common cause of isolated pancreatic agenesis. Integrating genome sequencing and epigenomic annotation in a disease-relevant cell type can thus uncover new noncoding elements underlying human development and disease.
@Mr. Michael
I was planning on suggesting the Museum of Jurassic Technology and the La Brea Tar Pits (the the tar tar pits) to PZ. PZ, I hope you’ll have some time to explore.
Sometimes. It’s the beginning of our Springtime, and can be quite lovely. I love February in SoCal.
One of the CFI-LA paid employees will give him a ride. SOP.
Jebus, you think I don’t know about cis regulatory sequences? They represent roughly 3-5% of the genome, not 80%. Telling me that polymerase occasionally binds to a particular stretch of DNA just tells me that biochemistry has a certain level of intrinsic noise — it does not help me spot regulatory sequences.
You don’t know what you’re talking about, and you’re really missing the whole point.
I’ve been to both the Museum of Jurassic Technology and the La Brea tar pits, and I even have t-shirts from both to prove it!
Oh, shoot. Well, the email I just sent does have a couple of other suggestions…
@chris61
The word “functional” in the context that ENCODE uses it has little resemblance beyond that it stuck to some things.
So there is some small ability to actually reference the thing you pretend to defend.
I have had this open since yesterday. This is the second thread I have seen you have conspicuously resistant to actually describing the thing you are going on about. You seem completely untrustworthy on this issue. One last time since I’m pretty sure you are basically in here to make a lot of noise and push the burden of proof towards everyone but the person actually defending ENCODE, yourself. What do you know about how they controlled for false positives? Given the nearly useless definition of functional they use I’m entitled to just leave you here looking aweful and walk away.
“Sticks to things” Works quite well. That is the common denominator here because other than binding, which may be specific or non-specific, we have no idea how many of the things in their data sets are real until someone actually does something called a “functional study” for the proteins. Seriously, google “functional study” and protein (with the quotes). You will see some meaningful uses of the word functional.
Of course it’s not equivalent! That is our whole point! Their use of the word “functional” is utterly inappropriate unless heavily qualified! You liked my “first pass putatively functional elements” before. If I don’t see them qualify “functional” in similar terms they get mockery. A mere “80% of the genome is functional” is completely inappropriate.
My apologies if this has been addressed already, but how accessible will this talk be to someone who has basically no science background (say, AP-level high school courses)?
The whole point of using a screening test is to winnow down an unmanageable number so that a more accurate but more expensive or time consuming gold standard test can be used without it being too costly. A “screening” test with a definition of “positive” that is so broad that it ensures a positive result rate of 80% is practically useless. You may as well have just not done the screen, not bothered with the definition, and just done the “gold standard” testing directly.
@curiosity #82
It will be accessible. CFI-LA lectures are geared towards laypeople.
@ 81 Brony
What we have here Brony is a failure to communicate and this one statement points to what I think may be at the heart of it. This statement exhibits a fundamentally flawed understanding of most of the genetics, genomics and biochemistry on which ENCODE is based. I don’t even know where I’d begin to correct it.