More Recent Comments

Tuesday, March 02, 2010

"Highly Evolved" and Favorite Kingdom

 
You can kill two birds with one stone by visiting Byte Size Biology and reading the posting on Highly Evolved. Hopefully you'll never use the term "highly evolved" once you understand that it reveals something about your understanding of evolution. That "something" is not complimentary.

While there, respond to the poll on your favorite kingdom of life. The results so far are very strange, I would have expected bacteria to be in the lead since bacteria are so much more diverse and interesting that the other choices.


26 comments :

Anonymous said...

Do I have to go all the way to Warsaw to 'respond to the pole', or is this some kind of Festivus event?

The Other Jim said...

Excellent!

Now will someone tackle the dreaded "X % homology" next?

DK said...

Pure semantics. Only people who have nothing better to do argue about precise meaning of the terms. Isn't it the case the philosophers and mathematicians realized the futility of it long ago? Isn't it the case that philosophers and mathematicians are generally smarter than biologists? (Just askin').

Since the word evolution literally means rolling out and implies some sort of vectorial progression, I have no problem stating that "living fossils" are less evolved than organisms that are comparable but whose evolutionary history contains many more step-wise changes. Highly evolved = more evolutionary recent. Humans may or may not be more evolved than sponges. In most cases, I suspect, we can't even tell one way or another(?)

Larry Moran said...

DK says,

Since the word evolution literally means rolling out and implies some sort of vectorial progression, I have no problem stating that "living fossils" are less evolved than organisms that are comparable but whose evolutionary history contains many more step-wise changes. Highly evolved = more evolutionary recent.

It makes me sad to hear you say that your flawed way of thinking about evolution is a "pure semantic" difference.

It's not. You are wrong. There's no such thing as a "living fossil." Every single living organism is the product of billions of years of evolution. Evolution happens all the time—it can't be stopped.

As for the second part of your statement—where you say "highly evolved = more evolutionary recent"—that too, is a flawed way of thinking because I doubt very much whether you ever refer to recently evolved species of bacteria as "highly evolved."

And if you do, the term is useless and can simply be replaced with "more evolutionary recent."

El PaleoFreak said...

Evolution is change (not all sorts of change, of course). More evolved means more changed. Evolution is a scientific fact and we know there are different rates of evolution that depend on multiple factors, so there is nothing intrinsically wrong in expressions like "more evolved" or even "highly evolved".

Anonymous said...

Now will someone tackle the dreaded "X % homology" next?

Every year I try to raise the distinction between a homologous relationship between sequences vs. percent sequence identity/conservation, but generally without any noticeable success. :/

The Other Jim said...

El PaleoFreak,

Say we take a model organism and expose it to EMS or some other mutagen. The offspring are typically quite ill, as a result of increased DNA mutation rates affecting important genes. These organisms have experienced a burst of "sequence evolution". Are they more "highly evolved" than their non-treated cousins?

"Highly evoloved" in the context discussed in the article is a mistake, similar to "lower organisms". These phrases are implying a judgment, where none is warranted (sponges versus Einstein in the article linked to).

My experience is that evolution is the most mis-understood scientific discipline by the general public. If we wish to increase the general level of evolutionary literacy, we need to stop using these silly phrases and communicate accurately.

El PaleoFreak said...

The other Jim,
According to a well stablished scientific definition for evolution, the treated population of your example would be more evolved than the controls. Perhaps not "highly evolved". How time would the experiment last? :o)
I repeat: you can use "highly evolved" or "more evolved" in an unscientific way. But yo can also use it correctly. And you can explain why some uses are wrong. But these expressions are not silly themselves.
I know two people who avoid the word "evolution". They think "evolution" is a silly term because it implies pre-determined "development" of species and a "Great Chain of Being". It's a similar nonsense. If you want to communicate science, explain the scientific words and expressions, don't blame or bane them.

Anonymous said...

I don't think "more highly evolved" is necessarily bad, although I understand the way the term can be misused. However, suppose we compare drug-resistant TB to drug-sensitive TB. These two organisms are not equivalently adapted to their environment. Are they still equivalently evolved?

El PaleoFreak said...

"Are they still equivalently evolved?"

We need to know what is the ancestral state. If the ancestral state is -for example- "resistant", then the "sensitive" bacteria are more evolved.
They could be both more evolved and less adapted to their environment.

DK said...

Larry wrote:
It's not. You are wrong. There's no such thing as a "living fossil."

I think you just proved my point. Insisting on a precise and particular meaning of terms eventually leads to absurdities. Like "no such thing as living fossil". If it looks like a duck, walks like a duck and quacks like a duck then it is just a duck and I don't care if it spent 20 million years in the same niche accumulating a bunch of compensatory mutations. And if during the same time some other duck transformed into a supersonic jet navigating underwater by echolocation then again I have no problem to say that the latter formerly-known-as-duck organism is more evolved. And so on. All terms are contextual.

The other Jim wrote:
Now will someone tackle the dreaded "X % homology" next?

More useless semantics. I see this homology thing so frequently that I sometimes wonder if people get orgasms out of pointing that 90% of others use the term "incorrectly". Yes, it is incorrect. Blah. But as long as everyone understand what is being talked about, the terms serves is purpose. It's not like homology is some sort of sacred thing that is given to us from above. Most of the time, we establish homology based on sequence similarity (and a lot of times we are wrong). Since the term homology itself refers to similarity ("sameness"), it is only natural to use the two interchangeably and only insist on binary usage in those rare cases where it really matters. When it comes to sequences, the homology puritans defend tautology.

Larry Moran said...

DK says,

More useless semantics. I see this homology thing so frequently that I sometimes wonder if people get orgasms out of pointing that 90% of others use the term "incorrectly". Yes, it is incorrect. Blah. But as long as everyone understand what is being talked about, the terms serves is purpose. It's not like homology is some sort of sacred thing that is given to us from above. Most of the time, we establish homology based on sequence similarity (and a lot of times we are wrong). Since the term homology itself refers to similarity ("sameness"), it is only natural to use the two interchangeably and only insist on binary usage in those rare cases where it really matters. When it comes to sequences, the homology puritans defend tautology.

Homology is a conclusion based on evidence. "Homology" is a word like "pregnant"—you either are, or you aren't.

Sometimes the evidence for homology is based on sequence similarity. If the sequences of two genes are 75% identical then it's reasonable to conclude that the genes are homologous. If the sequences are only 15% identical then it is not reasonable to conclude that they are homologous.

When you say that two genes are 75% homologous you are misusing the word "homology." In that particular case, it's just a demonstration of ignorance and the ultimate conclusion isn't affected because the genes are homologous.

When you say that two genes are 15% homologous you are demonstrating two different forms of ignorance. You don't understand what the word "homologous" means and, in addition, you are promoting a conclusion that is not justified—the genes are probably not homologous.

The really interesting cases arise when two genes are between 20-30% identical. In those cases it's very important to separate the evidence (e.g. 24% identical) from the conclusion (e.g. not homologous).

The idea that you could use the words "homology" and "similarity" (identity) as synonyms for some cases but not for others strikes me as quite illogical.

Just out of curiosity, where is the cutoff that defines "... those rare cases where it really matters"?

Larry Moran said...

El PaleoFreak says,

Evolution is a scientific fact and we know there are different rates of evolution that depend on multiple factors, so there is nothing intrinsically wrong in expressions like "more evolved" or even "highly evolved".

The fact that there's a molecular clock that ticks at an approximately constant rate (equal to the mutation rate) argues against your statement.

True, there may be brief bursts of adaptation that fix a certain number of alleles above the background, but identifying those bursts isn't as easy as you might imagine.

Over the long term, all species have evolved from a common ancestor at an approximately constant rate. That makes it very difficult to identify those few species that have acquired more changes.

Let's say we can identify such species. Imagine that there's good evidence for increased fixation of alleles in the lineage leading to Plasmodium falciparum (malaria). Would anyone be comfortable referring to this as a "highly evolved organism" with respect to, say, Homo sapiens, which isn't "highly evolved"?

El PaleoFreak said...

Larry, I think there is a big bulk of scientific evidence showing that rates of evolution vary a lot between species, between clades and even between environments. This is valid for character (phenotypic) evolution *and* for molecular evolution too. Also, molecular clocks tick at very different speeds.
"Comfortability" doesn't matter here. I'm not discussing feelings. Personally I have no problem with Plasmodium falciparum being more evolved than Homo sapiens, if it is true.
You are right that in most cases, it is difficult to identify what species has strictly evolved more.

DK said...

When you say that two genes are 15% homologous you are demonstrating two different forms of ignorance ... when two genes are between 20-30% identical ... it's very important to separate the evidence (e.g. 24% identical) from the conclusion (e.g. not homologous)

[All other points in a different post, a simple point here]

C'mon, Larry - are you seriously going to insist that there exists a non-arbitrary cut-off number for determining homology??? LOL. And since you are calling me ignorant, I'll return the favor because this hits rather close to my actual work: Bacterial actin homologs MreB and ParM have 15% and 11% sequence similarity with actin. (You can put it into your textbook).

Besides, the meaning of homology you insist upon only means common origin. More likely than not, everything living originated from the same primordial puddle, so here you go - everything is homologous to everything. What a valuable concept! :-)

DK said...

Larry said:
"Homology" is a word like "pregnant"—you either are, or you aren't.

You are completely missing the point. Plus, I use the word "binary" referring to a certain meaning of "homology" and you reply with explaining to me that it is binary. What is it? Do you really think this is a difficult concept or something?

No. My point is different - words can, do and should have more than one meaning. When someone says "% homology", 100% of listeners understand what is meant. That means that the word/term served its purpose well. That all there is to it - becuase everyone understands that saying "% of common origin" is nonsensical.(*)

You don't understand what the word "homologous" means

News: the word "homologous" means "sameness" and as such it is used to describe a lot of very different things in different contexts. There is no natural law that says it can not also be used synonymously with "% sequence identity". (In fact, as the OP hinted, it *is* commonly used this way).

The idea that you could use the words "homology" and "similarity" (identity) as synonyms for some cases but not for others strikes me as quite illogical.

In which cases English language (and probably most if not all languages) must strike you as very illogical. E.g., how come one can use the words "justice" and "judge" as synonyms in one situation but not one wouldn't use the word "justice" in a different context? Nearly every dictionary entry offers more examples... Sorry, I don't see the big mystery here.

Just out of curiosity, where is the cutoff that defines "... those rare cases where it really matters"?

Whenever one explicitly discusses phylogeny. Then 100% of population automatically understands that the word "homology" is being used in its binary meaning. In the world I know, that's a great minority of cases.

--
(*) Strictly speaking, whenever you consider more complex cases like gene fusions, saying "partial homology" makes perfect sense. And anything that's partial can in principle be quantified and expressed as %.

Larry Moran said...

DK says,

C'mon, Larry - are you seriously going to insist that there exists a non-arbitrary cut-off number for determining homology???

Nope, the cutoff is somewhat arbitrary. It depends a great deal on the individual scientist. Some scientists are more knowledgeable about this subject than others. In fact, as you point out, the misinformation that's been spread in the literature by misuse of the words "homology" and "similarity" has led to a situation where the majority of molecular biologists are confused.

There's a percent sequence identity that most people will recognize as a solid indication of hoomology—that's about 30%. There's a number that most people will recognize as being very close to what you might reasonably expect from any two random sequences—that number is about 12%. All numbers in between require additional evidence in order to support a conclusion of homology.

LOL. And since you are calling me ignorant, I'll return the favor because this hits rather close to my actual work: Bacterial actin homologs MreB and ParM have 15% and 11% sequence similarity with actin. (You can put it into your textbook).

MreB and ParM show some sequence similarity to actin. It's possible that MreB, ParM and actin share a common ancestor and there was rapid divergence during the first two billion years. The rate of divergence slowed considerably after that because all actin genes in eukaryotes are much more similar, as are all MreB and ParM genes.

It's also possible that the MreB/ParM and actin genes are not homologous and the similarity arose by convergence. As it turns out, HSP70 genes (N-terminal domain) also show about 15% sequence similarity to actin genes and the structures are quite similar. However, it's very unlikely that these genes are homologous since this would require a very different kind of evolution than what we currently understand. This is a good example of convergence.

It's OK for you to propose the *tentative* conclusion that MreB and actin are homologous but to state the this is a scientific fact is ... not scientific. That's why it's not going in the textbooks.

Larry Moran said...

El PaleoFreak says,

Larry, I think there is a big bulk of scientific evidence showing that rates of evolution vary a lot between species, between clades and even between environments. This is valid for character (phenotypic) evolution *and* for molecular evolution too.

Please quote those scientific articles that claim "a lot" of difference between the rates of evolution in different species. I don't believe it. Note that I'm not denying *some* differences—perhaps even two-fold differences in some lineages. What I'm challenging is that there are big enough differences to justify the term "living fossils" and your claim that evolution may have been practically non-existent in some lineages.

Also, molecular clocks tick at very different speeds.

They tick at different rates for different genes. That part has been well-understood for almost half a century. But when comparing sequences of the same orthologous genes from different species the molecular clock ticks at an approximately constant rate. That rate isn't exactly the same in every lineage for several different reasons that we could get into but it's close enough that we can construct pretty good vertical phylogenetic trees where all of the branches end fairly close to each other at the top of the tree.

Let's assume that we understand why there's a molecular clock. It's because the mutation rates in different species are approximately the same and because the vast majority of changes we see are due to fixation of nearly neutral alleles by random genetic drift.

Now, you claim that there can be big differences in the rate of evolution between lineages. What parameters do you think are changed? Do you think the equations of population genetics don't apply to some lineages or do you think there are significant changes in the rate of mutation?

If it's the latter, what is causing the changes in rate of mutation? Since the vast majority of mutations are caused by errors in replication, is it because DNA replication is more error prone in those species? Or is the increase in mutation rate due to sloppy repair enzymes that don't fix as many replication errors?

DK said...

Larry wrote:
It's also possible that the MreB/ParM and actin genes are not homologous ... similarity to HSP70 genes ... it's very unlikely that these genes are homologous ... This is a good example of convergence

I see. You are simply being stubborn and don't want to admit that you were wrong. Fine. Of course it is your right to think that these are all examples of convergent rather than divergent evolution. But your opinion goes against nearly all opinions ever voiced on this subject. Facts from the existing literature:

1. There is a total and unequivocal consensus that Mreb, ParM and FtsA are true homologs of actin.
2. There are many, many, many papers that conclude and/or directly state homologous relationships within the actin/sugar kinases/Hsp70 superfamily of proteins (complete with trees and all that). E.g., the popular Bork et al. (PNAS,1992,89:7290–7294) has been cited close to 500 times. Another example is Kabsch&Holmes article written after actin structure was solved: http://www.fasebj.org/cgi/reprint/9/2/167.pdf (quote from it below)
3. I am not aware of a *single* paper that explicitely states that the actin/hexokinase fold is a product of convergent evolution.

Personally, I don't really care one way or another. All I am saying is that your arbitrary cut-off is meaningless - and that the main reason for it is that if one looks deep enough, everything is homologous to everything!

A quote from Kabsch&Holnmes:
Although structural similarity, as observed between trypsin and subtiisin, can also be the result of convergent evolution of unrelated proteins, displaying the same optimal solution for a common mechanism, this similarity does not extend beyond the catalytic center. The common core structure of actin, Hsc7O, hexokinase, and glycerol kinase shown in Fig. 2 includes not only the ATP binding site but also a-helices packing against the /3-sheets and the domain connections.
Moreover, the striking similarity of subdomains I (Ia) and 3 (Ila) within the common core structure would be difficult to rationalize as an outcome of convergent evolution, invented several times independently. It appears more likely that actin, Hsc7O, hexokinase, glycerol kinase, and perhaps also the prokaryotic cell cycle proteins (FtsA, MreB, StbA) (52) and the two E. coil phosphatases (53), are the product of divergent evolution from a common ancestor. A plausible scenario of evolutionary steps leading to the observed family of proteins would originate from a single-domain molecule. Probably this ancestor had some affinity for ATP, similar to hexokinase in the open conformation where the nucleotide binds to only one of the domains (43). A dimeric molecule evolved from this ancestor by gene duplication and fusion with improved ATP binding between its two domains. The homodimer then diverged into the sugar kinases, Hsc7O, and actin by numerous mutations in each domain. Additional subdomains have been acquired that are needed for a variety of specialized functions of each protein. From the widely dissimilar amino acid sequences and the low mutation rates it appears likely that this divergence is a very ancient event that happened before eukaryotic evolution.

Larry Moran said...

Dk says,

1. There is a total and unequivocal consensus that Mreb, ParM and FtsA are true homologs of actin.
2. There are many, many, many papers that conclude and/or directly state homologous relationships within the actin/sugar kinases/Hsp70 superfamily of proteins (complete with trees and all that). E.g., the popular Bork et al. (PNAS,1992,89:7290–7294) has been cited close to 500 times. Another example is Kabsch&Holmes article written after actin structure was solved: http://www.fasebj.org/cgi/reprint/9/2/167.pdf (quote from it below)
3. I am not aware of a *single* paper that explicitely states that the actin/hexokinase fold is a product of convergent evolution.


Everything you say is correct. The overwhelming consensus among molecular biologists is that HSP70 and actin are homologous.

That does not mean they are correct.

Let's try a thought experiment to see how much you've really thought about this.

Here are two facts. All actin genes in eukaryotes are very similar. What that means is that since the origin of the actin gene there has been very little change over a period of about two billion years. The conclusion is that most amino acid substitutions are detrimental.

The HSP70 genes are among the the most highly conserved genes in all of biology. Prokaryotic and eukaryotic genes are about 50% identical and there are large stretches of sequence that are invariant. What this means is that for over three billion years there has been strong selection against amino acid substitutions in those genes.

Now, here's your task. Construct a scenario where the actin genes and the HSP70 genes arise from a common ancestor. The end result has to be two different genes (actin and HSP70) that now differ by so much that their homology is not certain while, at the same time, explaining why the two genes are so highly conserved.

If you think about it seriously, the problem boils down to the fact that when you reconstruct the primitive ancestor of each gene (actin and HSP70) they still have to differ by >80%.

Now I'm not saying that convergence is the only possibility. What I'm saying is that the misuse of the word "homology" has obscured that fact that the conclusion is tentative and not proven. The fact that a majority of scientists might have have jumped to a false conclusion is very troubling but even more troubling is the fact that they don't even realize there's a problem.

El PaleoFreak said...

Spending some seconds in google it seems, for example, that rDNa of planktonic foraminifera evolves 50 to 100 times faster than in some benthic foraminifera: http://mbe.oxfordjournals.org/cgi/content/abstract/14/5/498
I think you won't have any problem finding LOTS of papers showing evidences that evolution rates can vary substantially. As for mutation rates, they can also be very different between genes, chromosomes, sexes, species and between higher taxonomic ranks. My impression is that these differences are not well understood yet.
I have not claimed anything about living fossils or about evolution stopping in some lineages. It was not me.

Larry Moran said...

El PaleoFreak says,

Spending some seconds in google it seems, for example, that rDNa of planktonic foraminifera evolves 50 to 100 times faster than in some benthic foraminifera: http://mbe.oxfordjournals.org/cgi/content/abstract/14/5/498

Oh dear! Do you believe everything that's published in scientific journals? The idea that the evolution of ribosomal RNA could be 100X faster in some lineages strikes me as absurd. Do you really think it's possible?

I think you won't have any problem finding LOTS of papers showing evidences that evolution rates can vary substantially

I agree. There are lots of papers that make such claims. The question is not that they exist, it's whether you seriously believe them.

As for mutation rates, they can also be very different between genes, chromosomes, sexes, species and between higher taxonomic ranks.

I don't this this is correct. Note, that doesn't mean there haven't been claims. It just means that on balance after many years of reading the literature I dismiss such claims. In most cases contradictory claims have also been published.

My impression is that these differences are not well understood yet.

Does it ever trouble you that controversial publications make claims that don't fit with what we know? I admit that sometimes new discoveries are published that cause us to re-think our assumptions but in most cases results that aren't well understood are simply wrong.

I think you need to be more skeptical. Don't believe everything you read in the journals. Check it against what you already know and be skeptical of things that conflict with that knowledge—while always keeping open the possibility that your knowledge may be incorrect.

Remember that paradigm shifts don't happen every week in spite of what some authors might claim.

DK said...

Larry wrote:
Now, here's your task. Construct a scenario where the actin genes and the HSP70 genes arise from a common ancestor ... the problem boils down to the fact that when you reconstruct the primitive ancestor of each gene (actin and HSP70) they still have to differ by >80%.

I think Doolittle did it already. Because he was sure that actin and FtsA are homologs and he had issues with the protein first evolving fast and then very slowly, he proposed an RNA-based organism that was making actin-like proteins. The mutation rate in such creature would be tremendously higher.

I may be very naive but I don't share Doolittle's reservations. An ancient protein had freedom to mutate and diverge all it wants - as long as it folded and remained soluble. The mutants that failed to maintain an ancestral stable fold were misfolded junk and resulted in negative fitness. Over time, the ancestral sequence diverged into distinct pro-actin, pro-hexokinase, pro-Hsp70 and a bunch of pro-[things we can't recognize or things that were eventually lost later on]. Here is your 80% difference. Then things started to get complex enough and division of labor became not only possible but also beneficial. Each of these proteins took on a distinct function. And it all happened to click and work well - that's your proto-cell. Since these are all very early things - and we know that early things that are useful are almost never discarded/changed (genetic code, translation), there is suddenly a strong pressure against any change in basics. Making new things - sure, but don't dork with the foundation because everything else is built on top of it. Hence practically no more evolution. This logic holds true for other very ancient, little evolving families that have enormous amount of weak homologs across all kingdoms: GTPases, histones, ubiquitins.

It's time for you, Larry, to embrace the concept of "structural homology" :-) Just about everyone already did.

El PaleoFreak said...

First you say:

"Please quote those scientific articles that claim "a lot" of difference between the rates of evolution in different species"

Then, after a single reference (there are lots of them), you say:

"Oh dear! Do you believe everything that's published in scientific journals?"

I didn't expect this incoherent denialist attitude from you. It reminds me when a creationist ask for papers on the existence of transitional fossils, and then, when I give some to him, he say they are controversial and I should be more skeptical because I "believe everything that is published".

Paradigm shifts? Oh, dear...

El PaleoFreak said...

Spent some more time searching well documented differences in molecular evolution rates.
( One more example:
http://www.sciencemag.org/cgi/content/full/322/5898/86 )

Sorry, but I have to reassert myself. Although individual papers showing radical differences can be "controversial" (of course), the general evidence is overwhelming and it's of high scientific quality.
I can't see there is a real (scientific) "controversy" on this point.

In my opinion, a "pluralist" who is proud of accepting a variety of evolutionary modes and mechanisms should also take into account a diversity of rates, speeds and amounts of evolution. I can't see where is the "pluralism" in dismissing (via argument from incredulity) all this interesting differences in rates and their possible causes and factors. Fascinating stuff.

Larry Moran said...

Dk says,

It's time for you, Larry, to embrace the concept of "structural homology" :-) Just about everyone already did.

I embrace the concept of "structural similarity." In some cases, that structural similarity is due to homology but in other cases it's due to convergence.

The problem with most structural biologists is that they have embraced the concept that everything with a similar structure has to be related by descent. That's not science—that's religion.

You have put your finger on part of the exact problem we began with. If structural biologists would talk about "structural similarity" instead of "structural homology" they would see the flaws in their reasoning.

But they don't do that. As soon as they determine that, say, HSP70 and actin have a similar structure they immediately start referring to it as "structural homology." Thus, they skip the necessary step of assessing similarity to see if it's due to common descent or convergence.

That's bad science but it's become so common that most scientists don't even realize that they're making a mistake. Entire databases have been built on that false assumption (SCOP).