Chances Are Our Hearing Didn't Evolve "To Do" Anything

[Jud]

as an aside, it is always interesting to me when people try to discuss something in a technical area that is outside their training...

 

 

you 2 or 3 other other guys know who you aren't

 

 

Certainly.

 

Would you like to help out us non-experts by explaining what the genetic load issue indicates about the amount of "junk" in the human genome?

[Jud]

Depends upon what the gene product is. If it is a protein, for example, it may be able to tolerate a fairly high mutation rate, but if it is a structured RNA, or a regulatory RNA, perhaps much less so.

 

Do you believe this could result in no evidence of selection on a DNA sequence whose product may nonetheless be vital to survival and reproduction, in numbers sufficient to make up the difference between ENCODE's 80% and the ~9% of the genome that shows evidence of selection?

 

 

Not entirely sure what you mean

 

I asked Bill whether he considered evidence of selection on a DNA sequence to be a good indicator of "function" (I meant function as indicating the sequence played a role important to our survival and reproduction).

 

He replied that a protein coding gene could tolerate what he described as a "fairly high" mutation rate and still make a functional product (meaning, I presumed, that the sequence might not show up as being conserved, but its function nevertheless would be).

 

I then asked him whether he thought this would occur in such large numbers as to make up the difference between the ~9% of the genome shown to be under selection in various comparative genomic studies and the 80% of the genome described as "functional" (in the sense of having some biological activity, such as being transcribed) in the ENCODE study.

 

Bill's response (at least I took it as such) was along the lines that genomic studies done with commonly used techniques would be inherently flawed (to overestimate the amount of junk).

[Ralf11]

I would argue the two are the same.

 

or indistinguishable

 

'preserved' implies stabilizing selection; 'not cleared at the background rate of mutation' is unclear but could imply a weak selection gradient

 

 

Jud, I realize you are interested in this, but it is not something that can be explained over the internet, nor do I know where you are starting from. For some background reading, I'd suggest Wilson & Bossert's Primer of Population Biology, and various sections out of Wilson's Sociobiology.

[jabbr]

I asked Bill whether he considered evidence of selection on a DNA sequence to be a good indicator of "function" (I meant function as indicating the sequence played a role important to our survival and reproduction).

He replied that a protein coding gene could tolerate what he described as a "fairly high" mutation rate and still make a functional product (meaning, I presumed, that the sequence might not show up as being conserved, but its function nevertheless would be).

I then asked him whether he thought this would occur in such large numbers as to make up the difference between the ~9% of the genome shown to be under selection in various comparative genomic studies and the 80% of the genome described as "functional" (in the sense of having some biological activity, such as being transcribed) in the ENCODE study.

Bill's response (at least I took it as such) was along the lines that genomic studies done with commonly used techniques would be inherently flawed (to overestimate the amount of junk).

I think these issues are actually quite complicated. I'm not sure that the 80%/9% numbers are generally accepted.

 

Mutations are not randomly scattered in the genome and are known to localize in specific areas (see SNP). Most of the interest in these matters have to do with practical issues of looking at why certain people are susceptible to specific diseases or why certain drugs may cause side effects in specific people and have different efficacy in different people.

 

Computational genomics is its own subspecialty and I think the fellow whose article you quoted is involved in that area more than traditional molecular biology. Computational genomics tolerates a wider range of theories many of which do not share widespread acceptance among more traditional empirically based biologists. So called "Lamarkian" genetics has been widely discredited long ago, and the so-called "resurgence" that you are reading about really has to do with whether this ever exists, and while clearly not the vast majority of cases of inheritance, probably does occur in *very* specialized situations and when fairly global events occur that are able to affect the entire body including gonads e.g. perhaps fetal exposure to EtOH or toxins, or perhaps even neonatal exposure to breast milk etc, but even if that does occur on a very specific basis, would never extend to the giraffe example you mentioned. This is at best a fringe theory which probably is inappropriate for the lay public to delve into because the chance of getting it very wrong is vastly higher than the chance of gaining useful knowledge -- i.e. best to first learn the field of molecular biology.

 

Regarding epigenetics, I think the important questions are: OK you have N genes, how do these get turned on and off in specific cells/organs/individuals in response to either internal or external events.

Regarding mutation rates and the so-called genetic load: I think this is probably a bunch of BS because it doesn't take into account what can be very sophisticated gene repair mechanisms and redundancy which are applied selectively, etc. See those "silent" DNA regions have a whole lot more going on, doing a whole lot of things behind the scenes than we could ever imagine.

 

So lets say there are between 10,000 - 100,000 human genes. There are several orders of magnitude more individual loci of mutation (SNP) -- what does that tell you?

[Jud]

I think these issues are actually quite complicated. I'm not sure that the 80%/9% numbers are generally accepted.

 

 

Yep, they are certainly complicated and interesting. I would say the 80% number as an assessment of biological activity is pretty non-controversial. It was widely accepted before ENCODE that there was fairly pervasive transcription across much of the genome, for example. The controversial point was ENCODE equating this to its definition of "function." ENCODE has come in for a fair bit of criticism on that score.

 

There are multiple comparative genomic studies that have come up with the ~9% number. Similar to some of the issues Bill has raised, there has been discussion in the literature about technical challenges with sequence alignment across species, as well as various other issues. But if numbers discussed in some of these other studies are correct (20-30% of the genome possibly constrained by selection, several times 9% but still well short of the ENCODE number - again, ENCODE didn't purport to say 80% of the genome was constrained by selection, just that it showed some biological activity), we run into problems with mutational load and the molecular clock, as discussed below.

 

Computational genomics is its own subspecialty and I think the fellow whose article you quoted is involved in that area more than traditional molecular biology. Computational genomics tolerates a wider range of theories many of which do not share widespread acceptance among more traditional empirically based biologists. So called "Lamarkian" genetics has been widely discredited long ago, and the so-called "resurgence" that you are reading about really has to do with whether this ever exists, and while clearly not the vast majority of cases of inheritance, probably does occur in *very* specialized situations and when fairly global events occur that are able to affect the entire body including gonads e.g. perhaps fetal exposure to EtOH or toxins, or perhaps even neonatal exposure to breast milk etc, but even if that does occur on a very specific basis, would never extend to the giraffe example you mentioned. This is at best a fringe theory which probably is inappropriate for the lay public to delve into because the chance of getting it very wrong is vastly higher than the chance of gaining useful knowledge -- i.e. best to first learn the field of molecular biology.

 

 

Agreed.

 

Regarding mutation rates and the so-called genetic load: I think this is probably a bunch of BS because it doesn't take into account what can be very sophisticated gene repair mechanisms and redundancy which are applied selectively, etc. See those "silent" DNA regions have a whole lot more going on, doing a whole lot of things behind the scenes than we could ever imagine.

 

 

Sandwalk: Human mutation rates - what's the right number? discusses issues regarding the rate of mutation in humans. One problem is that at the end of the day you're constrained by the "molecular clock." If about an order of magnitude more of the genome is functional in the sense of affecting survival and reproduction than has been found in the comparative genomics studies (i.e., 80% vs. 9%), then to avoid mutational load problems the mutation rate slows down by 9 or 10 times, the molecular clock with it, and you've got the last common ancestor of humans and chimps walking around shortly after the Late Cretaceous. So the 9% figure seems to tie in better there.

 

 

So lets say there are between 10,000 - 100,000 human genes. There are several orders of magnitude more individual loci of mutation (SNP) -- what does that tell you?

 

 

The generally accepted number these days is about 19-20,000 protein-coding genes, in a genome size of about ~3+ billion bp.

 

The problem here is that whatever argument one wants to make should take into account that salamanders have genomes about 40 times the size of ours, and the domestic onion a genome about 5 times the size of ours. Either (1) salamanders and onions have much, much more complicated biochemical lives than we do (and some members of the Allium genus would have to have much more complex biochemistry than others, since there is a 4-fold variation in genome size among them); (2) our biochemistry varies at the most fundamental levels from living things that have larger genomes than we do, enabling us uniquely to "do more with less;" or (3) we share fundamental biochemistry with other life forms, and there's a fair degree of variation in the amount of "junk" macroscopic life forms carry around in their genomes (back to old-fashioned population genetics and the original article I cited again).

 

#3 seems to me to better meet POLA (principle of least astonishment), or to say it another way, extraordinary claims require extraordinary evidence, and what extraordinary evidence is there to support #1 or #2?

[Ralf11]

" the last common ancestor of humans and chimps walking around shortly after the Late Cretaceous"

 

?????

 

 

another error is the attempt to compare # genes in mammals vs. plants

[Jud]

" the last common ancestor of humans and chimps walking around shortly after the Late Cretaceous"

 

?????

 

 

another error is the attempt to compare # genes in mammals vs. plants

 

You know about the molecular clock, I'm sure. OK, slow it down by an order of magnitude. (Because you need a mutation rate about an order of magnitude lower than current estimates if you're not going to have mutational load problems with an 80% functional genome.)

 

Current estimates of the chimp/human line divergence using genetic methods (which would use the molecular clock) range from 5-7 million years, so let's take the middle number, 6 million. Now slow down that molecular clock by an order of magnitude, as explained in the paragraph above. This would put the chimp/human last common ancestor at 60 million years ago, creating a slight problem with the fossil record.

 

 

So there's a constraint on the rate of human mutation: It's got to correspond to a molecular clock that makes sense in terms of the fossil record.

 

 

Regarding plant and animal genes, throw out all cross-kingdom comparisons if you like. Compare plants within the genus Allium to other plants in the same genus. Or compare pufferfish to lungfish. Or salamanders or frogs to us. Why do some species have such a hell of a lot larger genomes than others? Are those huge genomes all 80% functional, and salamanders' biochemistry is fiendishly complicated? Or do genomes vary widely in the amount of junk they carry?

[jabbr]

Sandwalk: Human mutation rates - what's the right number? discusses issues regarding the rate of mutation in humans. One problem is that at the end of the day you're constrained by the "molecular clock."

 

Here you go again making assumptions. You are acting as though this so-called molecular clock is like an atomic clock in terms of precision. Who proved that? Even a cursory search turns up: Variation in the molecular clock of primates

 

If about an order of magnitude more of the genome is functional in the sense of affecting survival and reproduction than has been found in the comparative genomics studies (i.e., 80% vs. 9%), then to avoid mutational load problems the mutation rate slows down by 9 or 10 times, the molecular clock with it, and you've got the last common ancestor of humans and chimps walking around shortly after the Late Cretaceous. So the 9% figure seems to tie in better there.

again this "logic" if full of assumptions -- as above no reason to believe the "molecular clock" of humans and chimps is the same. You are trying to draw conclusions from pieces that just don't fit together.

The problem here is that whatever argument one wants to make should take into account that salamanders have genomes about 40 times the size of ours, and the domestic onion a genome about 5 times the size of ours.

 

I am reasonably certain that if onions can "hear" in some sense, that the genetics of their hearing shares very little with the genetics of our own hearing, and so whatever arguments are being made here do not need to take into account salamanders nor onions

 

That said, who thinks that the size of a genome directly determines the "complexity of its biochemical life". I can think of a number of reasons offhand why natural selection may not find the genome size to be the most important factor to optimize.

 

Let's get this back on topic: why might we not want to optimize the size of our music library in bits? Do you know of any algorithms which can losselessly compress the salamander genome? What about MP3? Would you be shocked if the salamander genome were to MP3 compress a factor of 40 more than the human?

 

If you start by assuming the scientific articles are god given facts, and then try to construct logical arguments on the basis of these articles, your conclusions can quickly get far astray of reality.

 

As my professor said on the first day of graduate school "Class, what I am going to teach you this semester is 50% wrong, the problem is that I don't know which 50%" A big part of science is dealing with uncertainty ... and no, we don't assume (3) because (1) and (2) seem extraordinary ... we look for (4) which is supported by evidence

[Ralf11]

some species have such a hell of a lot larger genomes than others in part because of duplication of genes

 

in part because of the fact that once you build a "big" organism - not in terms of size, but in terms of scope of its ability to deploy physiological or behavioral systems to deal with environmental exigencies - you no longer have to respond genetically to every problem the organisms face

 

I am biased as I work on such organisms, and there is likely a better way to state this, maybe using terms like euryhaline vs. stenohaline, etc.

 

I really think that adding some background would with understanding this area - it takes 2-3 years just to teach an undergraduate biology major the basic vocabulary we use, so it's not a completely trivial task

[Ralf11]

back on topic: to 'compress' a genome in something where size is at a premium, a really neat trick is to start different coding sequences with an initial offset and use the same DNA over & over again

[jabbr]

back on topic: to 'compress' a genome in something where size is at a premium, a really neat trick is to start different coding sequences with an initial offset and use the same DNA over & over again

 

[jabbr]

Salamanders are able to regrow limbs so perhaps they have huge LTRs: Activation of a LTR-retrotransposon by telomere erosion

[Jud]

Sorry, but you're spending some time here criticizing things I didn't say.

 

Here you go again making assumptions. You are acting as though this so-called molecular clock is like an atomic clock in terms of precision. Who proved that? Even a cursory search turns up: Variation in the molecular clock of primates

 

 

again this "logic" if full of assumptions -- as above no reason to believe the "molecular clock" of humans and chimps is the same. You are trying to draw conclusions from pieces that just don't fit together.

 

 

 

Think about this for a moment: I talked about a timeline from when the chimp and human lines *diverged* to modern humans. Where does that necessitate *combining* the chimp and human timelines?

 

 

I made no assumptions whatever about the molecular clock. I simply pointed out that a given background mutation rate will set a rough molecular clock; and that a molecular clock 10 times slower than the estimates currently used by scientists working in this field will be difficult to square with the fossil record. These are not assumptions I made or conclusions I drew; scientific discussions of how a given molecular clock rate will square with what we know of the development of the human lineage are common.

 

That said, who thinks that the size of a genome directly determines the "complexity of its biochemical life". I can think of a number of reasons offhand why natural selection may not find the genome size to be the most important factor to optimize.

 

 

Your second sentence is a decent summary of the original article I cited. The selection coefficient on genome size is very obviously not sufficient to uniformly limit it with the effective population sizes we get with macroscopic life. So it varies. The additional genomic material could be doing something significant to the organism's survival and reproduction, or it could simply be extra material that hasn't been gotten rid of because there isn't a strong enough selection coefficient relative to effective population size.

 

 

I invite you to have a look at the literature (you could start with publications that have cited ENCODE) to get a feel for the discussion about what percentage of our genome performs functions important to our survival and reproduction.

[Jud]

some species have such a hell of a lot larger genomes than others in part because of duplication of genes

 

in part because of the fact that once you build a "big" organism - not in terms of size, but in terms of scope of its ability to deploy physiological or behavioral systems to deal with environmental exigencies - you no longer have to respond genetically to every problem the organisms face

 

I am biased as I work on such organisms, and there is likely a better way to state this, maybe using terms like euryhaline vs. stenohaline, etc.

 

I really think that adding some background would with understanding this area - it takes 2-3 years just to teach an undergraduate biology major the basic vocabulary we use, so it's not a completely trivial task

 

Do you think variation in genome size is purely reflective of functionality affecting various organisms' reproduction and survival, or is some of it junk that hasn't been gotten rid of by purifying selection?

[Ralf11]

I don't know what "purifying selection" is.

 

I sure didn't teach it to my graduate students, but if I was still teaching it might be fun to toss it into a freshman course.

[Jud]

I don't know what "purifying selection" is.

 

I sure didn't teach it to my graduate students, but if I was still teaching it might be fun to toss it into a freshman course.

 

https://en.wikipedia.org/wiki/Negative_selection_(natural_selection)

 

 

Edit: Also https://scholar.google.com/scholar?q=genetics+purifying+selection&btnG=&hl=en&as_sdt=0%2C39&sciodt=0%2C39&cites=6704617216030606692&scipsc=

[jabbr]

I invite you to have a look at the literature (you could start with publications that have cited ENCODE) to get a feel for the discussion about what percentage of our genome performs functions important to our survival and reproduction.

 

There are two distinct categories of DNA, that whose function is known and that whose function is not yet known.

 

The genes which are responsible for peoples interest in reality TV is not important for our survival but might affect reproduction.

[jabbr]

Do you think variation in genome size is purely reflective of functionality affecting various organisms' reproduction and survival, or is some of it junk that hasn't been gotten rid of by purifying selection?

 

There's always hope... the problem is that the genes associated with interest in reality TV seems to be associated with a higher rate of early reproduction...

[Ralf11]

Jud, I don't rely on wikipedia tho it can sometimes help with background on non-technical areas.

[Jud]

Jud, I don't rely on wikipedia tho it can sometimes help with background on non-technical areas.

 

Yeah, I threw in Google Scholar too, but perhaps you wouldn't want to rely on it for something as technical as the meaning of the term "purifying selection."

[Ralf11]

I saw it - you might want to try being less of an internet expert in areas you aren't trained in.

[jabbr]

I invite you to have a look at the literature (you could start with publications that have cited ENCODE) to get a feel for the discussion about what percentage of our genome performs functions important to our survival and reproduction.

 

Do you understand the discrepancy in the 80% and 9% numbers you've quoted? Most if not all my comments are hints for you to see why the 9% conclusion is invalid -- to the extent that assuming a gene sequence with polymorphisms is less "important". Let me try to explain: suppose the gap has to due with massive redundancy of various types. The evolutionary value of such redundancy *could* allow the genome to survive despite a series of mutations, or it could have other value that you nor I can imagine.

 

Very plainly: the argument that because a gene sequence is not invariant over time, that it is not "important" is invalid. I find the papers that argue this to use trivial and frankly unsophisticated arguments. I've given you counterexample after counterexample so instead I invite you to go back and read every comment I've made and be sure you understand what I am saying. If you have a hard time understanding a specific comment and why it pertains to this argument, please ask and I'll try to explain.

 

Let's start here: suppose my music library is stored on a mirrored volume -- that means it takes up 2x the amount of space -- is that a waste? In the short terms, perhaps yes, but in the long term and statistically, the data is more protected. So I am paying upfront $$ for longer term protection. Evolution is a long term process, so while short term redundancy may result in large numbers of transcriptionally inactive genes, long term is different. -- and this is just one analogy, there are many many other issues.

 

But whether the ENCODE number is 80% or 50% or 95% is not really important in my view. The 9% number is totally irrelevant in my view. I've hardly begun to list the reasons why but think along the lines that mutations are encouraged and desirable in certain situations, that there is a strong evolutionary value in biodiversity. That things which appear to be inefficient in the short term turn out to be essential in the long term. That biology is smarter than we are. etc. etc. etc.

[jabbr]

Here's a good, brief introduction to SNPs: https://ghr.nlm.nih.gov/primer/genomicresearch/snp These "are the largest form of genetic variation among people". Now if someone assumes that areas of genetic polymorphism define areas of the genome that are not important "not important enough to be preserved" then you might be led to believe that these areas affect the so-called "junk DNA". Big big mistake. Turns out SNPs are very important for biodiversity: think different side effects of drugs, but also think: different susceptibility to things like Ebola ... so turns out really good to have in a population. Perhaps genetically "silent" until ... Ebola strikes or the Plague or something as yet undiscovered and then hmm...

 

Or more importantly for us here ... maybe SNPs determine whether we like tubes or SS, SET or PP, rock or classical, or perhaps SNPs determine which USB cables we prefer... who knows?

[Ralf11]

so, the polymorphisms have been purified enough??? ;]

[jabbr]

This review article addresses the mechanisms by which LINES and other repeats, which have been considered "junk" DNA, actually participate in the 3D structure of the interphase chromosome: http://medicine.yale.edu/lab/manuelidis/chromosomes/science_chroms_90_190065_284_23096.pdf

@wgscott noted the retrotransposons. These form LINES and SINES (above and https://www.ncbi.nlm.nih.gov/pubmed/10527415, https://www.ncbi.nlm.nih.gov/pubmed/7551548). These sequences alone form ~20% or more? of the human genome. I know nothing about the salamander nor onion genomes but wouldn't be surprised if they also have lots of lines...

 

Consider that certain sequences might not only be transcriptionally active *but also* have structural binding activity 'cause ya' know DNA is smart like that