More Recent Comments

Friday, May 01, 2015

Molecular Evolution Exam - April 2015

Here's the final exam in my course. Students have to answer the first two questions and three of the next five questions. How would you do?


  1. Choose a subtopic from your essay and explain it better than you did in your essay and/or rebut the comments and criticisms made by the marker/grader.

  2. Michael Lynch says in The Origins of Genome Architecture ....
    Nothing in Evolution Makes Sense Except in the Light of Population Genetics
    Evolution is a population genetic process governed by four fundamental forces, which jointly dictate the relative abilities of genotype variants to expand through a species. Darwin articulated a clear but informal description of one of those forces, selection (including natural and sexual selection), whose central role in the evolution of complex phenotypic traits is universally accepted, and for which an elaborate formal theory in terms of changing genotype frequencies now exists. The remaining three evolutionary forces, however, are non-adaptive in the sense that they are not the function of the fitness properties of individuals: mutation (broadly including insertions, deletions, and duplications) is the fundamental source of variation on which natural selection acts; recombination (including crossing-over and gene conversion) assorts variation within and among chromosomes; and random genetic drift insures that gene frequencies deviate a bit from generation to generation independently of other forces. Given the century of theoretical and empirical work devoted to the study of evolution, the only logical conclusion is that these four broad classes of mechanisms are, in fact, the only fundamental forces of evolution. Their relative intensity, directionality, and variation over time define the way in which evolution proceeds in a particular context.
    Do you agree with Lynch that “Nothing in Evolution Makes Sense Except in the Light of Population Genetics”? If so, why isn’t population genetics taught in introductory biology courses? If not, why not?

  3. Imagine that identical female twins were born to a woman in 1000 AD. Imagine that you could find a direct descendant of each twin in 2015. If you sequence the complete genomes of the descendants, approximately how many differences would you expect to find? How do these compare to the differences between any two randomly selected individuals from the same part of the world? Explain your reasoning and describe any assumptions you make. Think carefully before you answer. The second question is the most important one. (Human mutation rate = 130 mutations per generation. Haploid genome size = 3.2 × 109 bp.)

  4. Why do some scientists think that there is no unique tree of life?

  5. Many people believe that recombination evolved because it increases genetic variation in a population and this provided a selective advantage over species that didn’t have recombination. Do you agree with this explanation for the evolution of recombination? Why, or why not? What are the other possibilities?

  6. What is “evolvability ”and why could it be important in evolution? Why are some scientists skeptical of this claim?

  7. Richard Dawkins once wrote,
    Even the most ardent neutralist is quite happy to agree that natural selection is responsible for all adaptation. All he is saying is that most evolutionary change is not adaptation. He may well be right, although one school of geneticists would not agree. From the sidelines, my own hope is that the neutralists will win, because this will make it so much easier to work out evolutionary relationships and rates of evolution. Everybody on both sides agrees that neutral evolution cannot lead to adaptive improvement, for the simple reason that neutral evolution is, by definition, random, and adaptive improvement is, by definition, non-random. Once again, we have failed to find any alternative to Darwinian selection, as an explanation for the feature of life that distinguishes it from non-life, namely adaptive complexity.

    Richard Dawkins (1986) The Blind Watchmaker. p. 304
    Can you describe situations in Richard Lenski’s ongoing evolution experiment where neutral or deleterious alleles were essential for adaptive change?

131 comments :

colnago80 said...

Jerry Coyne thinks the same thing as Dawkins.

colnago80 said...

As an aside, Dawkins wrote that 30 years ago. Is he still as adamant today as he was then?

whimple said...

...Do you agree with this [selective advantage of increasing genetic variation over species that don't have recombination] explanation for the evolution of recombination? Why, or why not? What are the other possibilities?...

The explanation I like is that recombination evolved primarily to restore replication fork geometry following replication fork collapse in mitotic cells and secondarily to allow repair of DNA damage in mitotic cells by using either an undamaged sister chromatid or from a second parental homolog as a template for repair. So, the main evolutionary pressure for recombination is fitness of mitotic cells has little or nothing to do with genetic variation in populations. In mitotic cells recombination tends to reduce genetic variation by causing loss of heterozygosity.

There is a related question: given that organisms can use recombination to pair homologs thereby enabling meiosis and sexual reproduction, why is it that most organisms have in fact done that and reproduce sexually instead of asexually? (but that isn't the question that was asked)

Jonathan Badger said...

Which is particularly ironic given that his advisor was Dick Lewontin

AllanMiller said...

Many people believe that recombination evolved because it increases genetic variation in a population and this provided a selective advantage over species that didn’t have recombination. Do you agree with this explanation for the evolution of recombination? Why, or why not? What are the other possibilities?

As we are talking of 'species', the recombination we are presumably referring occurs in organisms that already have cyclic syngamy and reduction. The hypothesis that, in such an organism, recombinant genomes are on the average fitter than nonrecombinant genomes is at best a weak driver for the establishment of a complex, somewhat mutagenic system of DSB initiation and crossover resolution. Against a background of nonrecombiners, whose alleles are essentially whole independently segregating chromosomes (already potentially reaping a limited reward from genetic subdivision if n>1), a recombining locus may briefly prosper by attachment to fitter genomes, assuming the somewhat unlikely condition that the population is already sufficiently variant without it, but its immediate advantage dissipates when those genomes become common and variation diminishes. Its tendency to break apart positive epistatic interactions degrades its advantage further. Recombiners have little to gain from population effects while rare.

Any theory of recombination requires that a short term advantage permit its rise before its longer-term wider-scale effects can be felt. I think that recombination in the short term is a side-effect of the advantage accruing by crossover stabilisation of bivalents on the metaphase plate during reduction. In the long term, it gives greater evolutionary flexibility to lineages, and causes integration of advantageous traits in mosaic genomes with reduced selective interference between linked loci and more efficient adaptation, but these cannot be prime drivers for establishment.

Unknown said...

2. I agree with Lynch that nothing in evolution makes sense except in the light of population genetics. I also agree with the Society of Systematic Biologists that noting in evolution makes sense except in the light of phylogenetics. Most of all I agree that everybody should absolutely paraphrase Dobzhansky to make the point that their research is highly relevant. Nothing in evolution makes any sense except in the light of statistical paleontology (me, just now)!
There are a couple of subdisciplines one has to understand before getting anywhere near making sense of evolutionary theory as it stands. But it would be a mistake to confuse the vitality with singular centrality (and having written that I notice just how Gouldian a phrase that is). While population genetics is a neccessary component of any understanding of evolution, it is not itself sufficient. It's not taught in introductory courses, because apparently maths scares students.

3. Can not be decently adressed with the information given. Assuming 25yr generations, we are dealing with ~40 generations. We would need to know something about the diversity in the original population to estimate how many differences we expect simply through recombination and then we could compare this to that accumulated through novel mutations. It's likely that the genetic distance between the descendents of the twins is within the range of two randomly selected individuals, since after 40 generations they are close to being randomly selected individuals anyway (although that would depend on the population size).

4. Because they are idiots? Generally the argument is based on an ill-defined species concept, or it is based on the idea that horizontal gene transfer makes the notion of species moot, often in the latter case all incongruent gene trees are treated as evidence for HGT, even in cases where retention of plesiomophic states are a better explanation. In either case this means that what these people mean with "tree of life" is something that differs drastically from the notion of a tree of life that phylogenetics describes.

John Harshman said...

Can we agree that there are a few evolutionary processes that are not population genetic? I think species selection does happen, though it may not be all that important. And mass extinctions, as well as any other catastrophic events that happen rapidly with respect to organismal generation times, have little to do with population genetics.

Larry Moran said...

#2 B-
#3 B+
#4 F

Thanks for trying.

Larry Moran said...

Yes, we can all agree. Speciation, in general, is important but it has almost nothing to do with population genetics. Species sorting is not a population genetics phenomenon. Mass extinctions and other unique events are crucial for understanding the history of life although one could argue that those don't count as "evolution."

Unknown said...

I sense a disagreement on 4.
Maybe you would like to describe what the tree of life is, in your opinion. The appropriate species definition for phylogeneitcs is the internodal species concept and by definition this produces a tree topology (i.e. a graph with no trajectories from one vertex to itself that does not pass through at least one edge in two directions). Furthermore you can collapse any arbitrary graph into a tree with some simple operations. Doing this has always been a part of phylogenetics (if it wasn't things would go to hell as soon as you notice sexual reproduction). I haven't seen anything from the "there is no tree of life" side that wasn't predicated on misrepresenting what the tree of life is in the first place.

Unknown said...

Species selection is a population genetics process. It simply adresses populations of species, rather than populations of organisms...

Jonathan Badger said...

There are not very people left who still support the whole "There is no tree of life" extreme that was so popular in the early 2000s when HGT was a new thing. Even Koonin is back on the tree wagon although he calls it "the statistical tree of life" in his recent J. Mol. Evol paper.

The simple fact is HGT, while it certainly happens, isn't the rampant all-shuffling thing that was hyped up -- even in the microbial world, and even higher microbial higher order taxa exhibit quite distinct phenotypes as predicted by trees. Saying "there is no tree of life" is very misleading and basically only serves to fuel Creationist drivel about how Darwin was wrong.

Claudiu Bandea said...

Laurence A. Moran: Saturday, May 02, 2015 9:56:00 AM:

“#2 B-
#3 B+
#4 F
Thanks for trying.”


That’s ridiculous. Expecting students to give the answers you like is not a legitimate way of teaching science. If Simon gets a mediocre grade, what do you expect from your students, as it is unlike that they had the chance to accumulate Simon’s knowledge or his power of rationalization?

whimple said...

Recombination precedes and is required for meiotic synapsis, so "crossover stabilization of bivalents" considerably understates the role of recombination. Pairing of homologs prior to a reductive division requires a mechanism for comparing sequences of the homologs... that's what recombination does. Without recombination you don't get bivalents.

Unknown said...

"Speciation, in general, is important but it has almost nothing to do with population genetics."

In order to get speciation started, some genetic differentiation between subpopulations is needed. Population genetics tells us whether differentiation will arise, given values for gene flow, deme size, mutation rate, number of subpopulations, and differential selection across demes. Population genetics is absolutely central to understanding speciation.

Lou Jost

Rosie Redfield said...

#3:Approximately one base in 1000 would be different, just as for any two random individuals. After more than 1000 years (say 40 generations), the two genomes are very unlikely to contain any DNA segments that are identical by descent from this common ancestor. The mutation rate is irrelevant.

Larry Moran said...

My course is about critical thinking using scientific controversies as a vehicle to teach students how to argue.

When faced with a genuine scienttific controversy, it's absolutely crucial that you undestand the arguments of your opponents. I teach my students that they should be able to do a credible job of arguing both sides.

In order to get a good mark on Question #4, you need to correctly describe the views of those scientists who are skeptical about the existence of a universal tree of life. You cannot pass the quesstion by just giving me a bunch of reasons why you think there IS a universal tree of all life that most scientists agree on.

Larry Moran said...

@Claudiu

I suppose one could teach a course where students get full marks on an exam for writing whatever they want whether the instructor likes it or not.

You imply that this would be the proper way to teach science. I find this hard to believe but it does help me understand some of the things you say in your comments.

AllanMiller said...

OK, technically, a non-joined tetrad is not a bivalent. Nonetheless, the chromatid pairs still have to be hauled apart.

Not sure your proposed sequence is definitively true across all of life, ever. A quick Google suggests that even recently, the precise order had yet to be settled, though this 1998 paper argues for the sequence you suggest. http://www.nature.com/hdy/journal/v82/n1/full/6884870a.html

Regardless, modern synapsis and recombination have clearly co-evolved; I don't think we can necessarily assume that synapsis as we know it - the formation of the SC etc - has any bearing on the first organisms to do meiotic crossover. Recombination does do homology search, but then so does repair, from which recombination clearly derives. Something more coarse, in both reductive and nonreductive crossover pathways, still gets them close enough to do their fine search. Recombination's search is restricted to a very small fraction of the chromosome - a matching sequence to patch the DSB. I'm not sure that 'speculative' DSBs initiated in order to test broad homology would be a good idea! How do you patch the 'wrong' ones?

I'd presume one has to commence with an organism that has a syngamy-reduction cycle. It would be elaborate to propose that it had no mechanism for reduction until it had evolved crossover within it.. After all, if n=1, it doesn't really need homology search at all. Just haul 'em apart. But this runs the risk of aneuploidy due to mechanical difficulties; mobile crossover sites allow the chromosomes to be teased apart with even tension. When n>1, now we have a finer requirement, and it is possible that DSB initiation/resolution was the prime mechanism to enable synapsis, but I'd doubt it. It seems too easily fooled by mobile or incidentally homologous sequence to be trusted with this primary role.

I assume we agree, nonetheless, that the cytological role of recombination is central to its evolution?

AllanMiller said...

There are, I think, evolutionary processes for which population genetics is an insufficient tool, rather than that there are no population-genetic factors involved. Mass extinction is not wholly divorced from allele fixation - fixation is extinction for something.

Population considerations do not cease to apply when there is no gene flow - it is just that there is a different dynamic when there is.

AllanMiller said...

No, #4 is a tricky one, as phrased. "Because 58 out of 64 invariant codon positions in non-mitochondrial codes just aren't enough for some people" might be my equally smart-alec, F-baiting response to the question. Of course students tend to avoid smart-aleckery in exams, if they know what's good for 'em. Nonetheless, "summarise the evidence for and against a universal Tree of Life" might be better.

Some consider that there was no single LUCA, but that somehow the genetic assignments of one protein-coding species can be integrated with those of another. This, however, evades the question of how we can get two independently-origined, yet interconvertible, protein-coding species in the first place. Others regard the decreasing degree of simple congruence between different gene trees when proceeding towards the base of the tree (post-LUCA, pre-Domain) as evidence that HGT was vastly more important than vertical descent in actual history, and is not merely an artifact of accumulation of single HGT events and other degraders of the descent signal.

Donald Forsdyke said...
This comment has been removed by the author.
Donald Forsdyke said...

“neutral evolution is, by definition, random, and adaptive improvement is, by definition, non-random.”

So let’s adapt by driving west,
From Kingston to T-town.
Ontario metropolest,
Held in much renown.

Head towards the setting sun,
Tank is full, in seat am sat
Mark, set, go … I hear the gun,
Alas, I have a random flat!

True, I can still westward drive,
But oh, so very slow.
With such a tyre I’ll not survive,
Extinction head to toe.

With random tyre I am curst,
My anger’s near to brimiting.
So, first things, first things, first,
Correct what is rate-limiting,

Cope with unfair deal,
Out with heavy spanner.
Change that flunky wheel
Then follow Nature’s banner!

Larry Moran said...

In their answer, I expect my students to recognize that recombination arose long before eukaryotes and meiosis. They will lose marks if they spend any time at all discussing meiosis.

I also expect them to remember that recombination is not restricted to meiosis in eukaryotes. It occurs in all cells. Some of my students know that meiosis can occur in the absence of recombination.

Larry Moran said...

I would give students credit for smart-alec remarks if they were clever and correct. Yours is neither.

The genetic code may have a unique origin in some ancient cell but so do all genes that are universally present in all species. (There are only 100 or so genes that fall into this category.) We cannot be certain that all these genes, and the original genetic code, were present in a single last universal common ancestor (LUCA). That's a common assumption but it's not supported by phylogeny.

More importantly, as you point out, the question isn't about the existence of LUCA. It's about whether there's a unique tree of life that describes the main branches from hypothetical LUCA to all modern species. Your smart-alec comment would have revealed that you didn't understand the question so you would lose marks.

AllanMiller said...

Hah! You confused me by talking about 'species'. Perhaps you could more clearly distinguish to which of two completely different processes (albeit related in certain essentials) your question refers.

Recombination is certainly not restricted to meiosis, but the part that is principally responsible for variation in eukaryotes is. If you aren't interested in eukaryotes, perhaps you could say so.

AllanMiller said...

You don't find 'as you point out' and 'indicate that you don't understand the question' somewhat inconsistent? It's no good just taking part of my answer as the whole. "The question" being: "Why do some scientists think that there is no unique tree of life?". It is not a clear question, harbouring numerous potential viewpoints they may adopt. I lose marks for discussing the wrong one? I demand a recount!

Larry Moran said...

Lot's of people get confused about the evolutionary advantage of recombination amd how it evolved. We covered this in class so all of my students know that recombination is common in prokaryotes. Maybe you should take my course next year? :-)

Larry Moran said...

The question was perfectly clear and students were expected to describe several points of view of scientists who deny that there is a unique tree if life. You do not cover any of them.

Tom Mueller said...

Oh I hope I am not misleading my students in high school

I equate recombination with "sex" (based on previous conversations I realize that Allan may take umbrage)

So why is sex beneficial?

The initial benefit to sexual reproduction was to short-cut “Muller’s ratchet” E.g. two individuals each with one copy of some deleterious mutation will produce offspring that are free of either mutation 25% of the time. Furthermore, sexual reproduction permits organisms to shed themselves of deleterious gene combinations faster than by reverse mutation (the only asexual option available)

A later derived benefit of sexual recombination was to enhance parent’s chances of enhancing offspring' adaptation in order to survive Darwinian competition, remembering that each generation produces far more progeny than can possibly survive.

That would mean - sex/recombination did not start out as a “some selective advantage over species that didn’t have recombination." Recombination was an exaptation of DNA repair – a random and fortuitous side-benefit.

Now of course, I could open a different Pandora ’s Box by wondering out loud about “hierarchical levels of selection” i.e the species level perhaps.

I am just curious if my scribbles so far have passed muster.

Tom Mueller said...

post script @ Larry:

I have said on more than one occasion that I for one would love to sit in one of your courses!

Unknown said...

"It's about whether there's a unique tree of life that describes the main branches from hypothetical LUCA to all modern species."

I'm not sure what you mean by "main branches" here. The universal tree of life consists of internodal species. These by definition have a tree topology. The internal structure of an internodal species does not have a tree topology, but is a more general graph (which is what some people would refer to as a web structure). So when I come across statements that debate whether there's a TOL because "at the root it looks more like a web", I can't help but facepalm. It shows that the person making that statement is woefully ignorant of what the TOL even is. That there is a TOL follows tautologically from universal common descent. There's no way to have UCD and not have a tree topology that resolves some of the total topology. And we've known that it doesn't resolve all topology even before Darwin. Heck, there are web-like structures in the biblical genealogies and of course Darwin children could trace back to their great grandparents in two different ways, which makes the Darwin household a really good illustration of how non-tree topologies work in biology. No one ever claimed that the TOL resolves everything. It does resolve the relationship between internodal species, nothing more, nothing less. When you find evidence for a lof of gene flow between two sets of organisms, whether that's HGF or sexual reproduction you find evidence that they are one internodal species. That's an interesting result. But it's not a reason to reject the TOL.

By a similar token, I've heard that ENCODE has shown most of the genome to be functional. Of course that point rests on using the term functional in a way nobody else does, but it makes their results more spectacular and of course there was a splashy press release. The "there is no TOL" side was just as predicated on redefining fundamental terms to make their claims sexier and generate press coverage as that. There is no reasonable disagreement over the existence of the TOL among people who understand phylogenetic systematics.

Larry Moran said...

I think everyone has a pretty good idea what a tree looks like and a pretty good idea what a tree of life should look like if it's similar to a real tree. On an idealized tree of life you should be able to trace the lineage of every modern species by following a single path down the tree until you reach LUCA.

We know for a fact that this doesn't work with eukaryotes because at some point two large branches fuse to become a single branch. From that point on there are two separate pathways to LUCA. I'm talking about the main endosymbiotic event that gave rise to mitochondria.

There may be other fusion events in the evolution of eukaryotes (other that chloroplast). In addition, there are plenty of examples of small branches fusing to large branches and genes are passed from the tip of one growing branch to the tip of another.

The result is something that isn't a unique tree of life. It may be a net or a web but it may not even be a unique net or web. At this point we can't even rule out the possibility that life arose more than once although that seems unlikely.

Tom Mueller said...

So how did Simon's answer get an "F"?

I thought his rendition of HGT not too dissimilar to your "fusion events"

Both seem to mirror Doolittle's "uprooting the tree of life"

Tom Mueller said...

@Allan

Treading where Angels fear to tread:

I once suggested something along the lines that bacterial conjugation could be considered sex not coupled to reproduction

You disagreed. Would this be a good juncture to reopen that can of worms?

AllanMiller said...

Tom - if ya like! I got nowhere last time, so will probably get nowhere this. Larry is determined to pursue the notion that recombination is sex, and is 'for' variation. I simply disagree.

Recombinational enzymes are indeed closely related to DNA repair. That I do not dispute, nor the fact that prokaryotes integrate DNA in a manner commonly termed 'recombination'. What I don't agree with is lumping this all together as one thing, with common properties and function. It serves no particular purpose to call bacterial processes 'sex'. They are analogous at best, despite the fact that there are homologous proteins involved in the processes. It is particularly silly to call transformation and transduction 'sex'. The proteins of eukaryotic recombination have some homology with those of prokaryotic recombination, but that is in part because they share homology in turn with repair pathways. They serve a different purpose, and operate, as an ensemble, in a very different way.

It doesn't really matter too deeply what one calls things, but it can obscure differences to try and lump it all together. I don't even think the reciprocal recombination of eukaryotes is eukaryotic sex's most striking feature - that is serial alternation of ploidy. The questions 'why did prokaryotic recombination arise' and 'why did eukaryotic recombination arise' are separate. But it looks like I'd fail Larry's course for thinking so, and for making the reasonable assumption (not having done the course) that when one talks of 'species' one is talking of biological species, and hence of eukaryotes. Prokaryote recombination is not restricted to particular strains, but can take place across domains.

More interesting though, on the 'how did sex arise' (the eukaryotic kind, FFS!) question, is 'why did haploid organisms start to form binary units'? It's not just to make chimeric genomes.

Rosie Redfield said...

Larry, I anxiously await your judgment. Do I get an A?

Graham Jones said...

In phylogenetic analysis, you need to distinguish between species trees and gene trees. A species tree is a parameter in a model. A gene tree is a parameter in a different kind of model. HGT does not violate either model, it just changes the relationship between the two kinds of tree: the gene tree(s) no longer fits inside the species tree. Endosymbiosis on the other hand, does violates the model for a species tree, as do hybrid species on a smaller scale. (Recombination can violate a gene tree model, but that does not seem relevant to TOL discussions.)

Larry Moran said...

You have given the correct answer but you need a bit more explanation in order to get an "A." The students know that they are supposed to spend 30 minutes on each question and write something substantive. Almost all of them did that and almost all of them reached the correct conclusion.

I'm a little nervous about this question because it comes perilously close to being a "trick" question. A trick question is one where the instructor deliberately misdirects students so they will choose a wrong answer.

In this case, the students had all the questions three weeks before the exam so they had time to think about it. (They had a list of 28 questions and they knew that six of them would be on the exam.)

Larry Moran said...

In my couse we define the tree of life as a species tree. In order to get full marks on the question, the students have to demonstrate that they know the difference between a gene tree and a species tree. Thus, when discussing HGT they have to discuss whether the conflicts between different gene trees are real and how these relate to the species tree of life. They have to cover endosymbiosis because that's crucial. They have to discuss the various fusion hypotheses that may have given rise to eukaryotes. They have to cover the Martin & Graur paper on the tree of 1% and why that's important. They should know something about the difficulties of rooting trees and the difficulties of computing gene trees when dealing with deep phylogeny.

They do NOT have to agree with the opponents of the tree of life.

Larry Moran said...

I am totally opposed to the idea that sex and recombination are synonyms. However, we do discuss those papers that consider recombination to be the purpose of sex and whether mixis is equivalent to recombination. (It isn't.)

I'm totally opposed to the idea that sex or recombination evolved because it is an adaptation for increasing variation in a population. My students don't have to agree with me but the papers they have been assigned make a convincing case.

We discuss the Red Queen Hypothesis and whether this can explain the evolution of sex or just it's maintenance. My students understand the twofold cost of sex and Muller's ratchet. They understand that the adaptive advantage of eukaryotic sex is an unsolved mystery but the reason for recombination can be explained.

When talking about the evolution of sex we use yeast as our eukaryotic example and I make it vey clear to students that eukaryotic sex evoled in single cell organisms. That's what they need to keep in mind when thinking about this problem. They also have to keep in mind that prokarotes and eukaryotes share a common ancestor so you cannot have an intelligent discussion about the evolution of sex or recombination without involving bacteria.

Unknown said...

I think everyone has a pretty good idea what a tree looks like and a pretty good idea what a tree of life should look like if it's similar to a real tree. On an idealized tree of life you should be able to trace the lineage of every modern species by following a single path down the tree until you reach LUCA.

Well, I gave the technical definition above - a tree is a graph in which there are no paths from a node to the same node that do not run along at least one edge in two directions. That's equivalent to stating that it is loop-free. We could also note that we can be pretty certain that the TOL is strictly dichotomous, although there are scenarios in which it would generate true polytomies (but generally when you find a polytomy you just have insufficient data to resolve actual dichotomous branching).

And again I note that we can obtain a tree by collapsing a more general graph.

From that point on there are two separate pathways to LUCA. I'm talking about the main endosymbiotic event that gave rise to mitochondria.

I do think it is reasonable to treat mitochondria as endosymbionts and thus as separate from the eukaryotes. There is a population dynamic within the eukaryote cell and while we are now dealing with obligate symbiosis that's not a good reason to not place the mitochodria separately from the eukaryotes The TOL is about ancestor-descendent relationships, it's not about ecological interactions. If the latter were to be incorporated we would have to wonder which interactions would count and how. Obligate mutualism isn't that rare and is not that different from endosymbiosis. But just mutualism isn't that different from obligate mutualism. And once we are there, how about host specific parasitism?

But the key issue here is that we know a lot of non treelike graphs charting ancestry anyway. There is a pathway to LUCA from me through my mother and there is a pathway from me to LUCA through my father. That did not pose a challenge to the TOL, because the TOL was defined in such a way that these loops would be collapsed into internodal species. Graham above discusses hybrid species and the internodal species concept does not recognize "species" that hybridize as separate. You can always collapse these loops and end up with a tree. Now, we know this tree does not resolve everything, but it resolves some things. A legitimate debate can be had on how much of the history of life is resolved by the tree. But there is no legitimate debate on whether there is one.

Unknown said...

They have to cover the Martin & Graur paper on the tree of 1% and why that's important.

Do you mean the Dagan and Martin paper? Graur and Martin wrote a brilliant one on molecular clocks, but the tree of one percent paper was by Dagan and Martin.

Now, it all goes awry in the very first sentence of the abstract:

Two significant evolutionary processes are fundamentally not tree-like in nature - lateral gene transfer among prokaryotes and endosymbiotic gene transfer (from organelles) among eukaryotes.

Here's how that sentence should read:
Three significant evolutionary processes are fundamentally not tree-like in nature - lateral gene transfer among prokaryotes, endosymbiotic gene transfer (from organelles) among eukaryotes and sexual reproduction.

Of couse sexual reproduction is different from the other two in that it was known when the idea of a TOL was developed. It lacks novelty in this sense. But it also supplies a model of how not tree-like processes can be incoporated into the TOL with not much hassle. Nobody is arguing that non tree-like processes don't matter. But the TOL has been able to handle them from the get go.

Tom Mueller said...

'morning Larry

So what would your definition of sex? Is conjugation in Bacteria "sex"?

Tom Mueller said...

@ Allan

I wonder out loud whether transitory "merodiploidy" in bacteria represent a incipient bacterial version of "serial alternation of ploidy" although the point is moot.

I always presumed that "the reciprocal recombination of eukaryotes is [indeed] eukaryotic sex's most striking feature!! ...even though reciprocal recombination was initially some exaptation to DNA repair.

Tom Mueller said...

Hi John – Hi Larry

Re:
Speciation, in general, is important but it has almost nothing to do with population genetics. Species sorting is not a population genetics phenomenon.

Forgive my impetuous and persistent naïveté, but I am not so sure.

A venerable elder in the AP Bio community just this week flagged this excellent review:

http://www.humancond.org/_media/papers/crespi04_vicious_circles.pdf

Crespi cites ample evidence in the literature describing how positive feedback loops at a population level can accelerate species sorting and speciation.

ITMT – I thought Bacterial Conjugation represented sex decoupled from reproduction whereas the life cycle of the Ichneumon wasps would represent an example of decoupling meiosis from recombination addressing your point regarding “mixis”.

Maybe I am confused again. What am I getting wrong?

Tom Mueller said...

Let's try that again:

...the life cycle of the Ichneumon wasps would represent an example of where meiosis is decoupled from recombination until the supply of host caterpillars diminishes. At this juncture, the likelihood of one caterpillar being parasitized by more than one wasp transitorily re-couples meiosis to recombination.

So what is the definition of "sex" if not recombination?

judmarc said...

I'm totally opposed to the idea that sex or recombination evolved because it is an adaptation for increasing variation in a population. My students don't have to agree with me but the papers they have been assigned make a convincing case.

We discuss the Red Queen Hypothesis and whether this can explain the evolution of sex or just it's maintenance. My students understand the twofold cost of sex and Muller's ratchet. They understand that the adaptive advantage of eukaryotic sex is an unsolved mystery


Heck, it's just math ;-) : http://link.springer.com/article/10.1007%2Fs12064-009-0077-9

judmarc said...

In my course we define the tree of life as a species tree.

What would that look like in regions where the biological species definition doesn't apply?

AllanMiller said...

We discuss the Red Queen Hypothesis and whether this can explain the evolution of sex or just it's maintenance. My students understand the twofold cost of sex and Muller's ratchet.

Yeah, so do I. Do you discuss that the twofold cost of sex is considerably overhyped? Being, as it is, an issue only for organisms with asymmetric gametes, and therefore not a barrier to the evolution of sex at all. It may be a barrier to the evolution of those gender asymmetries, but first things first. Gender asymmetries arise against the background of an established sexual species - indeed, an established sexual ecology: an entire clade of sexual species, and probably only in multicellular forms. So sex can go a long way without ever encountering the twofold cost, even if it were a real cost.

It is much less clear in a realistic background, taking all factors into account - differential evolutionary rate, multiple variant loci, Hill-Roberston interference, the cost of gene conversion in a perennial diploid, etc - that gender is penalised at all by the twofold cost of males, simplistically accounted. Clones may occasionally extinguish the parental species, but the naive expectation is that this should be the rule, without much rigour to back that up.

They understand that the adaptive advantage of eukaryotic sex is an unsolved mystery

Probably because it does not really have one in simplistic population-genetic terms. You cannot rely on simplistic population genetics to investigate the very process that draws the boundaries of the sexual population upon which it depends. The dynamic across the boundary is significant - asexual reversion is instantaneous speciation, which isn't readily assimilated into the simple models. And mechanistic considerations also have a part to play. Which is why I might also disagree with the Lynch quote at this point.

but the reason for recombination can be explained.

Which one, eukaryotic or prokaryotic? And how?

AllanMiller said...

So what is the definition of "sex" if not recombination?

There is not a hint of reciprocal recombination in the colloquial usage. Even though technical usages often depart from the colloquial, I don't see the sense of removing all lay meaning. Sex is intercourse; children are made by sexual activity. Asexual reproduction would not be reproduction in which recombination was absent, but reproduction in which only one parent had input (automixis is on the boundary).

What would you call a life cycle that only had achiasmate meiosis, n=1? Sexual, surely? Two parents contribute to each zygote. It's combination, not recombination, that defines sex IMO. In bacteria, of course, one can then stretch the concept of 'parent' to breaking point and point to a scrap of DNA from a dead cell or virus as the second parent!

Larry Moran said...

Simon Gunkel

Yes, I meant the Dagan and Martin paper. I was responding from home on my iPad and relied on my memory, which is becoming increasingly unreliable.

Dagan, T. & Martin, W. (2006). The tree of one percent. Genome Biol 7, 118.[doi: 10.1186/gb-2006-7-10-118]

There are two other important papers that we covered.

Pace, N. R. (2009). Mapping the tree of life: progress and prospects. Microbiology and Molecular Biology Reviews 73, 565-57 [doi: 10.1128/​MMBR.00033-09 ]

Doolittle, W. F. (2009). The practice of classification and the theory of evolution, and what the demise of Charles Darwin's tree of life hypothesis means for both of them. Philosophical Transactions of the Royal Society B: Biological Sciences 364, 2221-2228. [doi: 10.1098/rstb.2009.0032]

Larry Moran said...

Simon Gunkel says,

A legitimate debate can be had on how much of the history of life is resolved by the tree. But there is no legitimate debate on whether there is one.

There's clearly a legitimate scientific debate over the existence of a unique tree of life. There are many prominent, intelligent, scientists and philosophers who argue that there is no tree of life. I asked my students to examine and understand those arguments.

YOU may think that there's a unique tree of life—and therefore no debate in your mind—but that's an opinion not a fact. You can't dismiss the existence of a debate just because you strongly favor one side.

At some point in time, scientists reach a massive consensus on most topics. In such cases we can legitimately say that there is no longer a scientific debate. There's no scientific debate over evolution, for example.

Larry Moran said...

Do you discuss that the twofold cost of sex is considerably overhyped?

No, but I probably should.

[concerning the reason for recombination]
Which one, eukaryotic or prokaryotic? And how?


We can understand and appreciate that recombination evolved as a mechanism of DNA repair. Sine recombination is universal, this explanation works for all species, prokaryotes and eukaryotes.

Tom Mueller said...

@ Allan

How would "achiasmate meiosis, n=1? " not be an instance of mitosis? Again, I must be confused. Could you describe Metaphase I?

Tom Mueller said...

@ Allan

This is like Alice through the looking glass.

I always considered serial brother-sister incest by parasitic haplodiploid mites and wasps as clonal expansion by Meiosis; raising interesting implications when considering the Red Queen Hypothesis

judmarc said...

Question re #4: How important is it whether there is a "unique" Tree of Life? With HGT/LGT, does it really matter which microbe happened to provide a particular gene or piece of a gene to which other microbe and eventually to which eukaryote whenever the event occurred historically, when precisely the same gene or piece of a gene could have been provided by a completely different microbe that happened to pick it up via HGT/LGT? In other words, is this not simply contingency, and how important is the fact that microbe A rather than microbe X was in "our lineage" (beyond the fact that contingency is undeniably a part of the history of species)? Or to put it a different way, isn't this piece of question #4 just question #3 writ large (i.e., after all this time, the identity of the specific ancestor becomes irrelevant to the result)?

AllanMiller said...

We can understand and appreciate that recombination evolved as a mechanism of DNA repair. Sine recombination is universal, this explanation works for all species, prokaryotes and eukaryotes.

Fair enough. HR is one mechanism of repair. The consequences of DSB resolution vary a lot though, and only some are recombinational in the sense of changing the genetic sequence (depending on the degree of sequence homology in the source of the patch, in non-end-joining pathways). Repair is certainly universal, with good reason, and some form of recombinational product is frequently inevitable, but I'd say this could well be a side-effect.

Reciprocal recombination in eukaryotes invokes the repair pathways by 'deliberate' initiation of DSBs, but the sister chromatid is not favoured as the source. Gene conversion products and crossover products are produced with about 50% impartiality. It is certainly possible that sufficient DSBs are initiated to create one crossover on average not for the reason of generating recombinant products, but for cytological stability. The sister chromatid would make a better patch, but would not perform the same cytological role, and hence we get reciprocal exchange as a byproduct, not the reason.

AllanMiller said...

How would "achiasmate meiosis, n=1? " not be an instance of mitosis? Again, I must be confused. Could you describe Metaphase I?

n is the haploid number, so in the diploid there are 2 chromosomes, one from each parent, achieved through syngamy instead of replication.

I think the fact that it 'looks like' mitosis could be relevant. If pre-sexual cells only ever did mitosis, with division triggered by the presence of duplicated chromosomes and an enlarged cell, an almost indistinguishable state could be provided by syngamy. It would mainly lack attachment at the centromere (assuming early mitosis had this feature). In this way, the earliest problem could be to keep the partners together (assuming a benefit to their coming together), rather than evolving a special means to haul them apart.

AllanMiller said...

I always considered serial brother-sister incest by parasitic haplodiploid mites and wasps as clonal expansion by Meiosis; raising interesting implications when considering the Red Queen Hypothesis

Yep, there are many variations on the theme! Not sure how much variation you can squeeze out of the same few chromosomes though.

Tom Mueller said...

Allan

You lost me. I simplify Mitosis vs. Meiosis for my students insinuating they merely understand what is happening at the metaphase plate in Mitosis vs. the two metaphase plates in Meiosis (without memorizing all the different phases)

Metaphase I is by definition reductional division. If n already = 1 , then reductional division is impossible.

On the subject of definition - if recombination and sex are NOT synonyms: I would be grateful if someone were to provide me a definition for "sex" acceptable to all Biologists, especially Larry.

Tom Mueller said...

damn autocorrect... I meant

I simplify Mitosis vs. Meiosis for my students INSISTING they merely understand what is happening at the metaphase plate in Mitosis vs. the two metaphase plates in Meiosis (without memorizing all the different phases)

AllanMiller said...

Tom, n=1 means gametes have 1 chromosome each, so the diploid complement is 2n. If there is a 2-step meiosis, you have 4n chromatids in metaphase I. I defined it minimally to eliminate independent segregation, a 'crossover-like' feature as far as genetic assemblage is concerned, but I probably confused rather than illustrated by doing so. 'Achiasmate meiosis' would have served.

But my point is, I don't think you would would call it 'asexual' simply because Prophase was achiasmate - ie, crossover does not seem to me to be a fundamental part of the definition of sexual reproduction, although it is the norm.

Whether acceptable to all - or even Larry! - or not, I would define sex as serial alternation of ploidy through syngamy and reduction (with a boundary case where both gametes sample the same prior diploid set). I'd reduce ambiguity by referring to 'crossover', 'gene conversion' and 'repair' when that's what I meant, and call those bacterial processes which involve alteration of genetic sequence by integration of 'foreign' DNA 'transformation', 'transduction' and 'conjugation' respectively. I would call the process of DSB repair by homologous recombination as 'repair by the HR pathway'.

Of course, until everyone accepts my nomenclature, every discussion of sex and recombination will be preceded by a lengthy discussion of the topic "but bacteria do it too!". ;)

Tom Mueller said...

Hi Allan

As I understand it, cross-overs between homologous chromatids are not instances of recombination.

Regarding conventional usage: Bacterial Conjugation is invariably called "sex".

ITMT, I wonder out loud whether the peculiarities of Ciliate conjugation fail your stringent criterion to be called "sex".

AllanMiller said...

As I understand it, cross-overs between homologous chromatids are not instances of recombination

I don't think that's right. Crossovers between sister chromatids might not be, but if crossovers between homologues aren't recombination, then it has no part to play in eukaryotic sex, which seems to contradict your attempt to synonymise.

Regarding conventional usage: Bacterial Conjugation is invariably called "sex"

I dare say - invariably wrongly! It just looks a bit like multicellular intercourse (the plasmid does not even pass through the pilus), and involves genetic material from 2 individuals. It is clearly analogous, not homologous. It's fair enough if people want to define it that way, but it confuses the hell out of discussions on the 'evolution of sex' when people do that - same with recombination.

When people talk of the 'evolution of sex', they almost invariably mean the evolution of the eukaryotic system. The roots of that system lie in prokaryotes, naturally enough. But eukaryotic sex is not apparently a derivative of conjugation. There is not a strong genetic or process continuity, simply a series of exaptations only some of which perform similar roles in the two systems, a natural consequence of the widespread requirement for homologous repair and other DNA management tools.

RAD51, for example, is homologous to bacterial RecA, and both proteins do homology (!) search and hold single and double strands together to assist strand invasion. In both cases, the proteins serve in multiple pathways that provide DSB repair. But only some of these are associated with sex, however defined. I think it more accurate to say that 'RAD51/RecA initiate DSB HR repair however caused' than 'RAD51/RecA are linked by their role in 'sex''.

It's a semantic point, but I think the popular usage causes confusion. The eukaryotic process of combination of a binary haploid organism into a single diploid and subequent reduction has a completely different dynamic - and really is a completely different 'thing' - than conjugation, common proteins notwithstanding. The proteins are held in common because HR repair of DSBs is, not because sex is. Whatever advantages accrue to conjugator or conjugee, they cannot be assumed relevant to the existence of eukaryotic sex. When eukaryotes used RAD51 to repair DSBs and form crossovers, they did so because RAD51 will leap into action whenever a DSB is detected, just as RecA will. RecA->RAD51 were conserved by their role in repair, not their role in 'sex'.

Arlin said...

I have to agree with Larry (while dissociating myself from his over-the-top no-one-is-right-but-me style). The vast majority of biologists have decided that the chimaeric nature of eukaryotes is an exception that can be ignored. They decided this before there was a debate about HGT. Simon, I respect your logical nature, but you want to redefine us and our mitochondria as 2 different species having a very intimate ecological relation, so that you can preserve the notion that there is one tree? Really? I think this just proves Larry's point.

Arlin said...

I would be happy to rebut the arguments of Lynch and Dawkins. They are both making totalizing claims, one for Darwinism and the other for "population genetics".

Dawkins draws on a dichotomy between selection and randomness. The whole argument rests on this dichotomy, which is meaningless. The same meaningless dichotomy is used by Darwinians to distinguish mutation and selection. Neutral evolution is "by definition" "random". Really? What is the definition of "random"? If "random" means "mechanistic" (as opposed to teleological, magical, or guided by the beneficent hand of the creator), then everything in scientific explanations is random. If "randomness" is intended to refer to any of the ordinary scientific meanings of the word "random" (uniform, patternless, spontaneous, indeterminate, etc), then the claim is wrong.

When faced with questions about what they mean by "random mutation", Darwinists try to prop up their position with a special "evolutionary" meaning of random, by which a process is random if it doesn't consistently favor things getting fitter, i.e., a thing is random if it isn't like natural selection. This would make Dawkins's argument a truism.

Larry Moran said...

I agree with Arlin's view of Dawkins while dissociating myself from his over-the-top no-one-is-right-but-me style. :-)

Tom Mueller said...

,Hi Allan,

My bad... of course I meant to say that cross-overs between sister chromatids do not represent recombination.

But that becomes the only remaining option in your n=1 scenario you describe above.

Bringing me back to my original contention: what is sex if not recombination?

How do you deny Bacteria have sex but Ciliates do? Because exchange in the former is not reciprocal?

Mikkel Rumraket Rasmussen said...

There really are an enormous amount of confusion about what randomness means in science. Even among scientists I run into multiple different meanings when reading their work and it is frustrating as hell as a layman trying to grasp a concept when it is either not clear how the term is used, or when it is defined, another scientist will protest that definition and use another instead.

It would be really nice if scientists could get their collective heads together and agree on a definition of random and then start using it consistently in all fields of science. But even if they can't, it would still be an enormous help if they could agree on a definition at least within biology.

When even scientists can't seem to agree about some of these terms, what hope is there for the general public?

Larry Moran said...

I am slowly working on a book titled "Evolution by Accident." It's my attempt to avoid the debate about what "random" means and yet convey the notion that evolution is not predictable, nor designed.

Arlin said...

I started to write a response to Lynch's passage but it got too long so I made it a blog here: http://www.molevol.org/why-the-four-fundamental-forces-view-is-mistaken/

AllanMiller said...

Tom, My bad... of course I meant to say that cross-overs between sister chromatids do not represent recombination.

They do and they don't - depends on your definiton. If Homologous Recombination means patching a DSB with material from a related chromosome, then this happens in mitosis and bacterial fission as well, using the recently replicated near-identical copy, just as in the sister case.

If, however, Recombination demands that DNA sequence must change, then only exchanges with a more distantly related homologue would count.

But that becomes the only remaining option in your n=1 scenario you describe above.

Not sure why you're not getting this ... ;) Forget n, n wasn't as important as the achiasmate bit. You can have achiasmate meiosis with any karyotype. Still, I'll try again: in n=1 organisms there are 2n chromosomes per cell in the interphase diploid, because n is the haploid chromosome number. It's the number of chromosomes per cell coming out of meiosis, not the number going in. The diploid 2 are replicated during pre-meiotic interphase to give 2 pairs of sisters, 4 chromatids in all. Crossovers may occur (1) not at all (2) between sisters (3) between homologues. 2 and 3 may both occur in the same prophase. How can (1) not be sex, while anything including (3) is? It makes no sense at all to exclude ploidy alternation from the definition, and instead make it all about something that may or may not occur in any given cycle.

How do you deny Bacteria have sex but Ciliates do? Because exchange in the former is not reciprocal?

Because there is no ploidy alternation in the former. I'll have to read up on ciliates, but they strike me as eccentrically sexual. The micronucleus is diploid; it gives rise to haploid cells which are exchanged and then fuse to generate new diploid micronuclei. Again, there does not need to be reciprocal recombination during ciliate meiosis to make it sexual reproduction - it would still be sexual if achiasmate. It's the ploidy that counts.

I think people sometimes get confused by the packaging, rather than remember the common roots of the process. See also: aphids, and other elaborate life cycles. They can all be broken down to the same basic cartoon:

haploid (with or without haploid mitosis) -> diploid (with or without diploid mitosis) -> haploid (with or without reciprocal recombination). This seems both natural and fully in accord with our general awareness of sexual life cycles in ourselves and in the living world. Recombination is cryptic, not universally performed, and should not be regarded as synonymous with or diagnostic of sex IMO.

Unknown said...

Well the alternative is to redefine the tree of life to preserve the notion that we and our endosymbionts are the same species. Endosymbiosis is a big enough deal to ensure that something has to change and treating mitochondria as separate from their hosts is far more consistent with how we treat other types of obligate mutualism (for instance in how lichen is treated). In the case of Cnidarians and dinoflagellates we even have cases of endosymbiosis, where we treat the symbiots as distinct. Since this relationship has likely started in the Ediacaran, we are dealing with a case where endosymbiosis lasting for more than half a billion years has not given us any reason to now treat the symbionts as the same species as the hosts. So rather than dismissing endosymbiont organelles as an exception that can be ignored, I think of it as analogous to the endosymbiosis of Cnidarians and the obligate symbiosis of lichen. It is a spectacular example for a process that has occurred quite a few times, but is no reason to topple the TOL. Phylogeny is not ecology.

YOU may think that there's a unique tree of life—and therefore no debate in your mind—but that's an opinion not a fact.

I'm not denying the existence of a debate, I'm denying it's legitimacy. In the 70s-80s there was a legitimate debate on where systematics would go, with a number of people arguing for phylogenetic systematics, some arguing for phenetics and some arguing for evolutionary systematics. The last group had the least tenable position, because Mayr and allies were trying to stay a middle course, but never got to clearly defining when they would favor phylogenetics over phenetics and vice versa. The debate ended sometime in the 80s with phylogenetics emerging as the only valid way to do things. Methods from phenetics still saw (and see) some use, because while they are not relevant for systematics, they are often relevant for ecological questions for instance.

Now, quite a few of the statistical techniques from phenetics got used in molecular systematics. There are some people who use them ignorant of their phenetic background and some people who recognize them as only approximately phylogenetic (it's worth noting that ML and Baysian techniques have moved away from this and some molecular data may even be better analyzed using classical phylogenetic approaches). Now, that debate also settled the TOL question. A phylogenetic scheme is always a tree.

This is basically where the arguments against a TOL fall apart. For instance the Doolittle article linked above states this:
Genealogy should always take precedence in a Darwinian classification (all true classification is genealogical), but most often there will not be a conflict with phenetics, basically because ‘like begets like’.
Now, the statement "all true classification is genealogical" settles things. It says: "There is a unique TOL!" But of course there's a second part to the sentence and it says: "And phenetics is often a good approximation to phylogenetics". So, further down we get this:
Traits most relevant to a practical and flexible phenetic or ecological classification have often been acquired by LGT, not through VD.
So now we are reviving a "phenetic or ecological classification", shortly after noting that it is not a true classification. At this point we have an argument that reads "There is a TOL if we use phylogenetics, which is the right thing to do, but if we use the wrong way to do systematics, we might be in trouble".
"I could not replicate the LHC results that show evidence for the Higgs boson by bashing my microwave with a baseball bat" is not a serious challenge to the standard model of cosmology and "if I do systematics incorrectly I don't recover the TOL" is not anything resembling a cogent argument against the TOL.

Unknown said...

Another gem from Doolittle is this one:
the proper way to model prokaryotic evolution over 4 Gyr is as a single, albeit highly structured, recombining population, not an asexual clade.

If only somebody had ever tried to figure out how evolution works in a single, albeit highly structured, recombining population. But it seems as if nobody ever did that. People have thought that all of evolution was down to asexual reproduction with no gene flow whatsoever. Maybe that's just trying to get as many biologists as possible to write margin notes saying "Sex!" as often as possible. It's certainly the paper that most begs for this margin note.
Alternatively one might feel compelled to write the following: There is a word for a highly structured recombining population and it is species. If the claim is that there is only one species of Procaryote we're talking! That claim might just be as provocative as the "there is no TOL" one and it is potentially correct as well. If there is a lot of gene flow through HGT then, yes, there's only one species of Procaryote.

Rosie Redfield said...

When starting any discussion of the evolutionary aspects of 'sex' and 'recombination', it's critical to begin by spelling out the definitions of these terms that the discussion should use. Otherwise we're all arguing at cross purposes.

Larry may not have needed to do this for his exam questions because he had already done it in his lectures.

Larry Moran said...

We discuss the evolution of sex but we take care to distinguish between various aspects of sexual reproduction. There is no standard definition of sex that a majority of scientists in the field would agree on. I make sure that my students understand this so they know why the problem of sex is so complicated.

Recombination is straightforward—the molecular mechanisms are taught in their molecular biology course and in their genetics course.

Jmac said...

Claudiu,

That’s ridiculous. Expecting students to give the answers you like is not a legitimate way of teaching science. If Simon gets a mediocre grade, what do you expect from your students, as it is unlike that they had the chance to accumulate Simon’s knowledge or his power of rationalization?"

Two of my kids have gone through UOT. They never complained. If you think Larry's system is not fair, it would be nice to let him that know first.

Tom Mueller said...

@ Allan

Do we agree that the independent segregation of chromosomes in gamete formation are also examples of recombination provided that 2n > 2

Which to my mind renders your "ploidy criterion" somewhat arbitrary

To my mind: "clonality" is the antonym of "sex"

If Meiosis in Haplodploid insects is an instance of clonal expansion without generating recombinants, then that instance of Meiosis would not be "sex"

If it is possible for discussions to proceed with different operational definitions of "species", can we please agree on one (and if necessary more than one operational) definition of "sex"?

Joe Felsenstein said...

The problem is, when most people hear the word "sex", they think you're going to talk to them about why the two sexes are different. Technically we are actually talking about outcrossing with recombination. But that doesn't sell books.

The word "sex" may be beyond redemption.

AllanMiller said...

Tom Do we agree that the independent segregation of chromosomes in gamete formation are also examples of recombination provided that 2n > 2

Which to my mind renders your "ploidy criterion" somewhat arbitrary


Aaaargh! This is why I used n=1 (2n=2) as my minimal case, to avoid the confusion introduced by independent segregation. But we seemed to be getting sidetracked by that, so I just focussed on achiasmate meiosis instead. Yes, we could call it recombination, albeit very distinct from the active process we have been referring to. IMO my "ploidy criterion" is less arbitrary than your recombination one, since it is independent of chromosome number and the presence of crossover.

When we are talking of the evolution of eukaryote sex, it is perfectly possible that n was indeed 1 initially - that the earliest diploids were formed from fusions and separations of mitotically competent haploids with 1 chromosome. Didn't have to be, but if it was, I would find it unsatisfactory to defer the appelation 'sex' until one of the chromosomes breaks and the break becomes common, at which point it starts to meet copies of itself and independent segregation commences, and suddenly it's sex because, well, recombination.

Of course n could have been >1 from the start. Either way, within a syngamy/reduction cycle that genome subdivision would provide (among other things) independent tuning of different haplotypes, a small step towards the more broad-scale generation of recombinant genomes proper. IOW, the bare bones of sex (regardless whether you want to lump in recombination or not) appear relatively easy to evolve.

The unavoidable nature of this segregational recombination is interesting, as is its precise 50% chance, mirrored in the more complex system by a 50% chance of crossover vs gene conversion. That is, meiosis has an internal symmetry from the start, and an impartiality that renders subversion by one partner or the other much more difficult.

@Joe - I tend to use the word 'gender' rather than 'sex'. 'Sex' I'd reserve for the sequential process encompassing both syngamy and reduction. Or reduction followed by syngamy if one prefers to put it that way round.

Tom Mueller said...

@ Allan

We are talking at cross-purposes.

For example the Bacterial Conjugation is mediated by the "Fertility" aka "SEX factor" which to your way of thinking apparently begs the entire question... to which I rebut that you seem to be arguing contrary to the conventional understanding of how Biologists understand the scientific meaning of the term.

Tom Mueller said...

@ Joe

re:
. Technically we are actually talking about outcrossing with recombination

Exactly what I was suggesting - but Larry specifically rejects any such equivocation of sex with recombination.

AllanMiller said...

Tom,

We are talking at cross-purposes.

Not entirely. Duelling definitions, maybe. What I am trying to argue against is this strange insistence that sex is defined by recombination. It doesn’t make any sense to me, particularly when one talks of early evolution. It predisposes one to miss a fairly obvious evolutionary sequence which ISTM has no inherent ‘costs’ whatsoever.

For example the Bacterial Conjugation is mediated by the "Fertility" aka "SEX factor" which to your way of thinking apparently begs the entire question

Indeed

... to which I rebut that you seem to be arguing contrary to the conventional understanding of how Biologists understand the scientific meaning of the term.

Stuff ‘em, then! Perhaps a small perceptual shift might help cut through the Gordian Knot of the Mystery of Sex (™) No-one is obliged to accept my definition, but I have been clear what it is and why. I have explained why I think ‘sex’ should be reserved for the eukaryotic syngamous alternation of ploidy, and why recombination is a poor alternative. It has nothing to do with sex in the eyes of the average speaker of English, therefore it invites misunderstanding, even among experts. “Do you mean sex sex or ‘sex’ sex?” Recombination does not occur to every chromosome in every cycle – and in some, it never occurs at all. Whereas ploidy always alternates.

Of course, ploidy is a somewhat alien concept as well, but we all understand what’s going on at the moment of fertilisation pretty well. Try telling someone ‘that’s not diagnostic of sex’, and instead point to crossover prior to reduction and say ‘… but that is!”. We're not obliged to consider lay meanings, but if we use a word, such connotations matter. Mine fits with lay understanding and points up the central symmetry of the process - a win-win!

AllanMiller said...

Joe . Technically we are actually talking about outcrossing with recombination

Tom: Exactly what I was suggesting - but Larry specifically rejects any such equivocation of sex with recombination.


Me too, though perhaps for different reasons. Not sure why Larry cleaves to the notion of ‘prokaryote sex’ though. But I would still disagree with Joe. Recombination is optional, even if near-universal. We need not even be talking of outcrossing in every generation. But we are always talking of serial ploidy change. So IMO, to understand sex, you have to start there. And if you do – if you envisage a haploid organism indulging periodic syngamous diploidy followed by reduction – where are the costs? Does sex suddenly become costly when you invent recombination? Mating types? Anisogamy? Multicellularity and dioecy? Not obviously, no. I think the proposed costs of sex are illusory from this perspective, as this basic system becomes progressively tuned by its own dynamics.

colnago80 said...

This is far off topic but what the heck happened in Alberta yesterday? Have they become the Peoples Republic of Alberta to join the Peoples Republic of British Columbia?

AllanMiller said...

The point of this exercise in definology is not simply to be pedantic. If one views recombination as 'the' reason for sex (which synonymising predisposes one to do), and that can't happen until you have syngamy and reduction, and can't have much population effect until widespread, then you have an awful lot of explaining to do. Syngamy is a process, so is diploid maintenance, so is reduction, so are crossover initiation and resolution. Hence, of course, the mystery of sex, but much of the mystery stems from some misplaced adaptive expectations that don't work too well when one considers the boundary of a 'proto-sexual' system. "It must be adaptive, otherwise it would not be so widespread", is the expectation.

Absent an offsetting benefit, it is expected to die out because of competition from perennial diploids, but they really aren't the threat they may seem. Yet even the paradigm of adaptation breaks down at the sex/no-sex boundary, and inappropriate choice of models (as even luminaries such as Maynard Smith and Williams were wont to do) muddies the waters further. Seeing permanent asexuality as a potential adaptation for a sexual species is of the 'not even wrong' school.

Petrushka said...

To my limited understanding, neutral implies purifying selection. Lensky implies that in a population having an assotment of neutral alleles, some may mutate into something having a selectable advantage. Over time, selection would be detectable, even if most change is neutral.

peer said...

I will reiterate:

"Nothing in biology makes sense except in the light of reproduction..."

The only thing that matters isn biology is NOT evolution, but reproduction. This should be elementary knowledge to every biologist.

From the above first law of biology, we can can understand all biology.

peer said...

"Can you describe situations in Richard Lenski’s ongoing evolution experiment where neutral or deleterious alleles were essential for adaptive change?"

The answer is:

50 x 10^3 (generations) x 50 x 10^9 individuals (per generation) = 2500 x 10^12 = 2.5 x 10^14 = 2.500 000.000.000.000 evo experiments to accrue 3 selectable mutations after drifting the pool for thousands of generations.

Whether or not neutral mutations play a role in Lenskis experiment is completely irrelevant for evolutionary processes, because such numbers (trillions, quadrillions) cannot be invoked for for instance the unique HARs in humans. From the putative ancestors to Homo less than 5 x 10^5 (generations) x (10^4) (eff pop size) = 5 x 10^9 = 500.000.000 is by far insufficient to select the accumulated 18 mutations present in human HAR1F after floating around for a few milions of years.

It is never gonna happen, Moran. Never. Except of course in Fairy Tale Land also known as "Evobrains".





Tom Mueller said...

Hi Allan


I always considered independent segregation of maternal and paternal chromosomes at Metaphase I of Meiosis as an example of recombination separate as separate and distinct from “chiasmatic” recombination, but recombination nonetheless.

Mitotic cross-over during between non-sister chromatids never constituted recombination in my mind because allele combinations remain constant (assuming no gene conversion) in the nucleus AND there is no “out-crossing” as Joe so eloquently phrases it.

Regarding Bacteria, I quote:

Sequence related families of genes and proteins are common in bacterial genomes. In Escherichia coli they constitute over half of the genome. The presence of families and superfamilies of proteins suggest a history of gene duplication and divergence during evolution. http://www.biologydirect.com/content/4/1/46

Of course, it gets better – exact duplicate copies of a lot of this Genetic information can exist on one or more free-floating plasmids separate from the main membrane-attached chromosome providing the Genetic material available for “outcrossing with recombination” reminiscent of incipient "ploidy". Microbiologists even have a term for it: "Merodiploid".

So what exactly are we describing in Bacteria if not “sex”?

judmarc said...

It is never gonna happen, Moran. Never.

Hey, you might wanna argue with this guy named Peer Torborg, who's been posting here arguing that evolution happens so frantically fast we know the Earth can't be old...oh, wait.

peer said...

Like everybody here, Rosie was wrong too.

A careful reader had noticed that Prof Moran explicitly stated that the "complete genome was sequenced". Here, like good ole NeoDarwinaisn e veryone tacitly assumd that SNPs only, and yes then they would be right. But guess what....the 1000 genomes project showed that the SNPS make up only the minor part of the differences between humans (0.1%). The real difference was found in the indel mutations (adding to 300 billion base pairs), which made up about 12 % of the genome.

The genome is extremely plastic and because the major part of genes is redundant anyway, losses and duplications are easily generated and mediated by repetitive sequences ("junk DNA").

In this "junk" lies the the origin of variation. Huge blocks (of frontloaded genetic information) can be missed or can be duplicated in order to spawn variation. Most of the rep seqs (TEs, ALUs, etc) qualify therefore as variation-inducing genetic elements (VIGEs).

peer said...

An accumulation of SNPs drives evolution NOT. It is the outdated NeoDarwinan saw, but it is dead wrong.

It is just the reshuffling of preexsiting genetic information and configuration of the chromosomes that does the evolution. No new information is required. Loss of info goes hand in hand with speciation. That is the conclusion of the huge genome projects.

Al last biology has become holistic. And we live to see it. Better open your eyes. It shows that most we thought we knew about everything is wrong.

Evolution can only be a fast process; it is an intrinsic property of genomes and selectable mutations will simply lead to an accrual of slighly deleterious. Completely in accord with the holistic data.

AllanMiller said...

Whereas genome rearrangements are never problematic ...

Tom Mueller said...

Maybe I am missing something:

Premise 1: Evolution can be described as a sequence of events, a history of life or genealogy as it were.

A trivial example would could be: “What happened first? The endosymbiosis event leading to mitochondria or the endosymbiosis event leading to chloroplasts. Of course, the Devil is in the Details. Ascertaining the correct sequence of events with accuracy may prove problematic, if not in this instance but in others. Either way, there is only one correct answer, either historical events happened one way or the other, not both.

Premise 2: If one subscribes to Doolittle's Genealogy should always take precedence in a Darwinian classification (all true classification is genealogical)...

The conclusion becomes simultaneously trivial and inescapable: There is in fact only ONE TOL! Figuring it out becomes the problem.

Now I am a big fan of Doolittle’s take on classification. Is there really any debate on that second premise?

So what exactly am I missing here?

Tom Mueller said...

@ Judmarc

I wondered about that. Why doesn't his head explode?

Larry Moran said...

About 82% of all ancient eukaryotic genes are more closely related to eubacterial genes than to archaebacterial genes. About 15% are closer to archaebacteria and 3% are ties.

Many, but probably not all, of the eubacterial genes come from the endosymbiotic event that led to mitochondria.

Your task, should you choose to accept it, is to draw a TREE that accurately reflects those facts.

Larry Moran said...

His head is very thick.

AllanMiller said...

Hi Tom,

I always considered independent segregation of maternal and paternal chromosomes at Metaphase I of Meiosis as an example of recombination separate as separate and distinct from “chiasmatic” recombination, but recombination nonetheless.

Sure, it is a form of genetic recombination, although with no linkage of the units, so any particular combination's persistence is somewhat limited.

And it is interesting; I do think that this kind of 'free' recombination had a part to play in early evolution of sex - ie, it is not necessary to have complex crossover to have at least some of the features of a recombinant system at the start - tuning of subgenome haplotypes, ie chromosomes. But note also that, crypically, you are referring to alternation of ploidy in trying to defend your 'recombinant' stance! Independent segregation does not occur unless there is a cycle of ploidy alternation. Which is rather my basic point. And it does not occur at all where the diploid set is 2, so I win! ;)

Mitotic cross-over during between non-sister chromatids never constituted recombination in my mind because allele combinations remain constant (assuming no gene conversion) in the nucleus AND there is no “out-crossing” as Joe so eloquently phrases it.

You keep getting your sisters and homologues (non-sisters) mixed up! No matter; if reciprocal HR occurs in the germ line, it can generate recombinant chromosomes, though this is probably a much lesser source than meiosis. Worth mentioning though that even with non-sister crossover (ie your recombination) 2 of the 4 products of 2 step meiosis for any given chromosome are nonrecombinant. If (as happens in ciliates and human females) 3 of the haploids are discarded, this means that (if n=1) "Tom-sex" only occurs in 50% of meioses ...

Regarding Bacteria, I quote:
[...]
Of course, it gets better – exact duplicate copies of a lot of this Genetic information can exist on one or more free-floating plasmids separate from the main membrane-attached chromosome providing the Genetic material available for “outcrossing with recombination” reminiscent of incipient "ploidy". Microbiologists even have a term for it: "Merodiploid".

So what exactly are we describing in Bacteria if not “sex”?


Something 'reminiscent' of it, by your own words! Analogy and homology. Might as well say viruses are having sex with their hosts. Now, about the dynamics of 'true' ploidy alternation ... ;)

AllanMiller said...

Tom,

Consider a mitotically competent haploid population in which an allele arose whose bearers were predisposed to fuse cytoplasm with another (for some reason!). This proto-diploid state could not be sustained indefinitely (I can give you various plausible reasons for that) but proved beneficial (ditto) and so caused increase in the representation of this allele in the population. Now, leaving aside the promissory notes in parenthesis, would we be justified in calling that system primitively sexual? Does it have much to do with what bacteria do?

Jmac said...

It is hilarious to watch you all usually very vocal and now speechless, and resorting to insults rather than science that is supposedly bulletproof.

@judmarc, Peer said it many times on this blog that he doesn't know how old the Earth is. He sympathizes with creationists because they are the only ones he came across that are willing to change their dogmas, unlike Darwinists. I personally don't agree with his approach, but I respect his choice.

judmarc said...

and resorting to insults

I said nothing insulting at all. I simply pointed out that in order to argue against evolution, Peer completely contradicts himself, saying in one response that evolution proceeds at a tremendously fast pace (his example of lake fish evolution), in another saying it can't possibly proceed fast enough for humans to have evolved.

He sympathizes with creationists because they are the only ones he came across that are willing to change their dogmas

Now *that's* funny!

Unknown said...

Your task, should you choose to accept it, is to draw a TREE that accurately reflects those facts.

This tree (if there was one) would not accurately reflect genealogy. You are overloading the tree with demands (that are not in fact demands that were originally made for the TOL) and of course at some point you find nothing that satisfies them anymore. But that is besides the point. The TOL is a phylogenetic concept it rests on genealogy and genealogy alone. Since the phylogenetic tree is already unique, at best additional demands are met by it. Alternatively no tree can do it. But it's worth noting that if we were to switch to a net structure to account for HGT, we would simultaneously move to a structure that did not reflect genealogy anymore.

AllanMiller said...

Petrushka - no, neutral = NO selection.

Tom Mueller said...

Hi Allan,

We are jumping back and forth here.

Re
You keep getting your sisters and homologues (non-sisters) mixed up!

Not this time! If there is no “outcrossing” it matters not whether or not crossing-over took place between sister or non-sister chromatids.

Joe rescued my original fuzzy and more naïve version. (Parenthetically - my sincerest thanks to one and all for your patience and your indulgence in bringing me up to speed)

I should have been more explicit when I endorsed Joe’s clarification above. To my mind: “Sex equals recombination with outcrossing!” The outcrossing was always implicit to my way of thinking but should have been clearly stated from the outset.

I still maintain that outcrossing does not require strict adherence to “alternation of ploidy” for us to be talking about “sex”. Imposing any such stricture would be most arbitrary. Your only defense is that science jargon should not deviate too far from a common layman vernacular which I counter: begs the question.

If a layman shout were to shout at his child: “Shut the door! Don’t let the cold in!”

… I for one would not urge a rewrite of the Laws of Thermodynamics to correspond more to more common usage.

Tom Mueller said...

@judmarc & Larry

Both of you: Stop it! You are killing me!

LOL! LOL! LOL! LOL! LOL! LOL! LOL!

AllanMiller said...

We are jumping back and forth here.

That's discussion for ya! ;)

You keep getting your sisters and homologues (non-sisters) mixed up!

Not this time! If there is no “outcrossing” it matters not whether or not crossing-over took place between sister or non-sister chromatids.


True, but that's not quite what was said. Let it pass. But note: there can only be outcrossing if there is alternation of ploidy. You can't be in the recombining population unless you are a producer of gametes, however many mitoses you do in the meantime.

I should have been more explicit when I endorsed Joe’s clarification above. To my mind: “Sex equals recombination with outcrossing!” The outcrossing was always implicit to my way of thinking but should have been clearly stated from the outset.

And so you are saying, seemingly without being prepared to admit the fact, that eukaryotic sex is fundamentally alternation of ploidy (albeit with arbitrary provisos that former haploid components of a diploid union must not re-establish that same diploid union, and either n>1 or there is crossover). Because without ploidy alternation, how do you get recombination or, for that matter, outcrossing, both of which depend entirely upon it (in eukaryotes)?

I still maintain that outcrossing does not require strict adherence to “alternation of ploidy” for us to be talking about “sex”. Imposing any such stricture would be most arbitrary.

It is not arbitrary at all. It is the universal feature of eukaryotic sexual systems. Other systems are, at best, parasexual, if we allow that 'sex' is used in biology precisely because it came from English, where it means neither 'recombining' nor 'outcrossing'. Admittedly, it doesn't really mean ploidy alternation either, but it does mean biparental inheritance, and genome fragments do not count (unless viruses are parents).

Your only defense is that science jargon should not deviate too far from a common layman vernacular which I counter: begs the question.

Hardly my only defence! I use that point to illustrate that we started out with a word, 'sex', which, it turned out when we started to look down microscopes, involved the precisely symmetrical fusion of half the chromosome complement of one parent with half that of another. Then we noticed that there were these little X's formed during reduction, and they corresponded with the linkage behaviour we observed in breeding experiments. And then, for no strongly defensible reason, people started to insist that that was sex, and ploidy alternation was incidental, because - look! - bacteria 'conjugate' and one squirts DNA into the other! So it made some sense to include the intercourse-like behaviour of certain bacteria, which is fine as far as it goes, but tells us next to nothing about the evolution of eukaryotic sexual systems. Most involve no intercourse-like behaviour at all, and are marked not by unidirectional transfer, but symmetry, at least genetically.

Recombination is of course a dramatic feature of the entire sexual eukaryote world, and forms the basis of most theories of the 'reason' for sex. I think this is an error. Ploidy alternation is not incidental, it is fundamental. The symmetry of eukaryote sex both drives it and keeps it stable. The symmetry is often cryptic, but it is real and fundamental.

If you want to understand eukaryotic sex (call the other what you like), start there.

Tom Mueller said...

@ Allan

Re:
You can't be in the recombining population unless you are a producer of gametes…

According to your definition yes – but therein lies the rub.

re:
… look! - bacteria 'conjugate' and one squirts DNA into the other! So it made some sense to include the intercourse-like behaviour of certain bacteria, which is fine as far as it goes, but tells us next to nothing about the evolution of eukaryotic sexual systems.

You mean like conjugation in Spirogyra where one cells squirts its DNA into the other cell through a tube just like Bacteria? No gametes here as far as I can discern.

I like to think that your version of "alternation of polidy" emerged later on as some exaptation to DNA repair, followed by the subsequent exaptation of incomplete Recombination with Outcrossing. In other words, "alternation of polidy" was not at all fundamental to "Recombination with Outcrossing" in its first incarnation when the transfer of Genetic Information at that point was beneficial but not necessarily complete.

Claudiu Bandea said...

Hi Tom,

Many thanks to you, Allan Miller, Simon Gunkel and some of the other contributors on this very interesting and meaningful discussion about the Tree of Life (TOL) and about sexual reproduction/recombination, and of course thanks to our host Larry of bringing forward yet another stimulating issue.

Here I want to second you on your revealing statement about TOL: ”There is in fact only ONE TOL! Figuring it out becomes the problem.”.

A few years ago, I made a similar point in context on a novel model on the origin and evolution of cellular and viral lineages: http://precedings.nature.com/documents/3886/version/1)

“The intent of the TOL, however, is to establish the line of descent among groups of organisms, or species, not necessarily the evolutionary relationships among their genes. Certainly, each of the millions of cellular and viral genes has an evolutionary history that can be revealed by a sequence-based phylogenic tree, but many of these gene-based trees do not represent a TOL that reflects the line of descent among the species. The problem, therefore, might not be with the TOL but with the reductionist approach of generating a TOL based exclusively on sequence-based phylogenetic analysis.”

Tom Mueller said...

Hi Claudiu

Welcome back!

... and damn! How uncanny is that! You and I just cross-posted on two different threads at the same time (actually 3 minutes apart)

http://sandwalk.blogspot.com/2015/05/ford-doolittle-talks-about-tree-of-life.html?showComment=1431199062588#c574695092488696898

AllanMiller said...

Tom,

Me: You can't be in the recombining population unless you are a producer of gametes…

According to your definition yes – but therein lies the rub.


You were talking of mitosis, and hence of eukaryotes. Are there any ways in which eukaryotes recombine at a population level other than through generation of haploid chromosome sets from diploids? And I include Spirogyra. I accept that my reference to 'gametes' may be misleading. I meant haploid sets. Ploidy alternation does not demand separate cellular packaging.

I like to think that your version of "alternation of polidy" emerged later on as some exaptation to DNA repair, followed by the subsequent exaptation of incomplete Recombination with Outcrossing.

This is very vague. The scenario I am proposing is precise. I don't think it is possible to be 'partially sexual', ie to have anything like the eukaryotic system without ploidy alternation. Sure you get HR in repair, but that's repair. You don't need to go to another individual to gain such repair - you possess duplicated chromosomes for a substantial part of the mitotic cyle. And you certainly don't need to swap genes to effect repair.

I don't think you can get from a genetic-fragment system to a precisely symmetrical one in 'Darwinistic' stages. If there is any fundamental asymmetry, one partner will tend to dominate - look at endosymbiosis. That said, I think that eukaryotic sex has more in common with endosymbiosis than anything else. The fundamental organism is haploid. When it forms diploids, this is a binary organism - a temporary union of 2 haploids to mutual benefit. But because (I argue) the origin of the system lay in the temporary fusions of 2 such haploid organisms, all diploids so created would be symmetrical from Day 1. Those diploids have since developed the idea that it's all about them. ;)

In other words, "alternation of polidy" was not at all fundamental to "Recombination with Outcrossing" in its first incarnation when the transfer of Genetic Information at that point was beneficial but not necessarily complete.

You seem determined to preserve a mystery of sex by sticking with a sequence which has little evolutionary logic. At some point, you have to explain ploidy alternation and diploid symmetry, because they are universal, and are the sole source of both outcrossing and recombination in eukaryotes. Why not start with it? And why be in such a rush to get this system recombining?

AllanMiller said...

[...]sole source of both outcrossing and recombination[...]

the recombination that is relevant to outcrossing, I mean ...

Tom Mueller said...

@ Allan

Ploidy alternation does not demand separate cellular packaging.

Interesting – I wonder out loud whether or not your thesis is “evolving”. That latest caveat may now accommodate Ciliate Conjugation in your scheme of things. Hmmm… Ciliates still never really ever have a “haploid stage” in the strict sense of the word.

http://users.rcn.com/jkimball.ma.ultranet/BiologyPages/C/Ciliates.html

Consider a mitotically competent haploid population in which an allele arose whose bearers were predisposed to fuse cytoplasm with another (for some reason!). This proto-diploid state could not be sustained indefinitely (I can give you various plausible reasons for that) but proved beneficial (ditto) and so caused increase in the representation of this allele in the population. Now, leaving aside the promissory notes in parenthesis, would we be justified in calling that system primitively sexual? Does it have much to do with what bacteria do?

http://sandwalk.blogspot.com/2015/05/molecular-evolution-exam-april-2015.html?showComment=1431097819737#c495423123193030002

Are you not describing the zygospore in the life cycles of “primitive” (OK I know I should not be using that term) Eukaryotes such as Spirogyra and Rhizopus?

You seem determined to preserve a mystery of sex by sticking with a sequence which has little evolutionary logic. At some point, you have to explain ploidy alternation and diploid symmetry, because they are universal, and are the sole source of both outcrossing and recombination in eukaryotes. Why not start with it? And why be in such a rush to get this system recombining?

I think right there you have distilled our debate down to its essential differences. There you have it.

I am going to need to think about that a bit. My intuitions presumed that the complexity of Meisosis would require intermediate more primitive transitional stages along the lines of "Recombination with Outcrossing" which initially conferred intermediate selective advantage but which was later supplanted by your alternation of ploidy which conferred even greater selective advantage.

Claudiu Bandea said...

Allan, Tom,

I find your discussion on sex/recombination one of the most meaningful scientific exchanges I ever had the chance to read here at Sandwalk; you might want expand on it and package it into an independent paper.

As you have already eluded to, it is critical to explore ‘sexual reproduction’ (and I emphasize *reproduction*here) and ‘recombination’ in an evolutionary context, specifically by addressing in more depth their putative evolutionary origins. Addressing 'why', 'how', and 'when' these processes originated should reflect on our perception and definition of their nature.

For example, in my model on the origin and evolution of cellular and viral “domains”, I pushed “sexual reproduction” (obviously, as a ‘primitive sex’ first; i.e. ‘pre-cellular and cellular fusion/fission events’) as the key process in understanding not only the evolution of the LUCA lineage and its descendants, but at the origin of life itself: http://precedings.nature.com/documents/3888/version/1

AllanMiller said...

Tom,

Claudiu's probably right, a paper might be better than blog-commentary!

Me: Ploidy alternation does not demand separate cellular packaging.

Tom: Interesting – I wonder out loud whether or not your thesis is “evolving.


I’ve not changed my conceptual tune, but recognise that I do have to choose my words carefully, for their inevitable baggage. There is for example a point at which the term ‘haploid’ shades into ‘gamete’, and it would be careless to talk of my original haploid population as a population of ‘gametes’. There are many variations on the ploidy-alternation theme relating to the enclosure of those haploid sets within cytoplasmic and nuclear membranes, as well as the presence or absence of mitosis in one or the other phase. But these are all derived states. Biological nomenclature was coined on the derived state.

Are you not describing the zygospore in the life cycles of “primitive” (OK I know I should not be using that term) Eukaryotes such as Spirogyra and Rhizopus?

Certainly if the life cycle involves free-living mitotically competent haploids which conjoin to make temporary (and potentially non-mitotic) diploids, that would be much the same thing. The modern states are derived, however, so we can’t say this bears more than a superficial resemblance to the ‘true’ primitive state I propose.

My intuitions presumed that the complexity of Meisosis would require intermediate more primitive transitional stages [...]

Your instincts were correct, I think. But the complexities you’d need going your way are multiple – you don’t just need to get diploidy and meiosis up and running, but the complexities of syngamy, requiring an improbable degree of simultaneity in their origin.

Here’s the $64,000 question: where did diploidy come from? There are only two options: syngamy and endomitosis. If you plump for the latter, you have no rationale for separation. The endomitotic diploid has no reason to yield haploids – and no-one but itself to merge them with. There is only a limited time during which the chromosomes will remain sisters (barring extensive gene conversion to keep ‘em in line), so if you envisage a lengthy intervening period of evolution, the haploids wouldn’t be haploids at all, and this divergence would counter the already weak net advantage of keeping a spare through G1.

Start with syngamy, however, and you can use the pre-existing mitotic apparatus to effect separation. In fact, it may initially be difficult to stop it. Two cells’ worth of cytoplasm plus homologous chromosome pairs will look, to cellular machinery, awfully like it’s reached the end of the S phase, barring the absence of a centromere.

[…] along the lines of "Recombination with Outcrossing" which initially conferred intermediate selective advantage but which was later supplanted by your alternation of ploidy which conferred even greater selective advantage

In both instances, why? What’s advantageous about recombination with outcrossing? And to what entity – gene, haploid, diploid, population or lineage? Perennial diploid lineages that don’t do it tend to have a short evolutionary lifespan (yes, rotifers, I know), but that’s not the same as saying R&O are advantageous in themselves. Organisms certainly don’t do it in order to avoid having a reduced evolutionary lifespan.

Tom Mueller said...

Hi Allan

Here’s the $64,000 question: where did diploidy come from? There are only two options: syngamy and endomitosis. If you plump for the latter, you have no rationale for separation.

Actually, ever since Jack Szostak’s seminal work on Double Strand Gap repair, this line of inquiry has flip—flopped in my head.

I am now more bewildered by how tetrads can be avoided during diploid Mitosis than by how tetrads are formed and resolved during Meiosis. You see where I am going with this?

What’s advantageous about recombination with outcrossing?

At the individual level? There isn’t any! Initially, at the individual level it was always only about DNA repair which only later represented an exaptation to what you already described as a proto-diploid state [which initially] could not be sustained indefinitely .

http://sandwalk.blogspot.ca/2015/05/molecular-evolution-exam-april-2015.html?showComment=1431097819737#c495423123193030002

This of course raises the complexity of our inquiry to a whole new level by answering Larry’s question #5 but yet leaving the door open on question #6 to further speculation whether our discussion may invoke one of those very rare instances where “species/clade selection” is indeed operative.

Anonymous said...

Allan, you wrote, "Two cells’ worth of cytoplasm plus homologous chromosome pairs will look, to cellular machinery, awfully like it’s reached the end of the S phase, barring the absence of a centromere."

As you know, there are organisms that do meiosis and mitosis without a centromere. This pattern is seen in such clearly derived organisms (e.g. sedges, wasps) that it must be derived there. It never occurred to me that it might also be the more ancient pattern, but maybe so. Centromeres don't have to have evolved with/before mitosis.

(This discussion has been very interesting and I've learned things. Thanks to both you and Tom.)

Tom Mueller said...

@ bwilson295

Judmarc cited an excellent reference above

http://sandwalk.blogspot.ca/2015/05/molecular-evolution-exam-april-2015.html?showComment=1430753085212#c8032607595907373748

The mathematical models consider three reproduction pathways: (1) Asexual reproduction, (2) self-fertilization, and (3) sexual reproduction. We also consider two forms of genome organization. In the first case, we assume that the genome consists of two multi-gene chromosomes, whereas in the second case, we consider the opposite extreme and assume that each gene defines a separate chromosome, which we call the multi-chromosome genome.

That bizarre second case extreme scenario is an accurate description of the Ciliate macronucleus as described in the Kimball link I cited above…

http://users.rcn.com/jkimball.ma.ultranet/BiologyPages/C/Ciliates.html

…except those chromosomes broken up into pieces ("minichromosomes") that often are so small that they contain only a single gene occur in the macronucleus (not the Mininucleus apparently).

I remain baffled how mitosis proceeds under these circumstances and remain grateful to Allan for prodding me to investigate these lines of conjecture further.

AllanMiller said...

Thanks Barbara, that's pretty interesting. I had not wondered particularly if the ancestral state had centromeres in mitosis or not - it's not vital either way, since there are mechanisms that cope with their absence even in normally centromerically-attached sisters. But it would certainly not hurt the scenario to speculate that there may have been none!

AllanMiller said...

Tom,

I am now more bewildered by how tetrads can be avoided during diploid Mitosis than by how tetrads are formed and resolved during Meiosis. You see where I am going with this?

Well, this is an interesting aspect. Maynard Smith (in Evolution of Sex and Major Transitions) wondered why, if the diploid state provided any advantage at all, it ever reduced back to haploids. Some theories suggest a cyclic switch in the relative advantage of the 2 states, but there are at least two mechanistic possibilities that don't involve an adaptative explanation. I've mentioned one: the possibility that the partners would be hauled apart by mitotic mechanisms almost as soon as they got together. The first task would therefore be to prolong the diploid state. But another, not mutually exclusive possibility is that primitive mitosis simply could not proceed in the diploid - that reduction had to happen in order for replication to occur. It can't even be taken for granted that mitosis could handle more than 1 chromosome pair. This kind of thing leads me to disagree with Lynch. Some things in evolution make sense because they are mechanistic limitations.

Me: What’s advantageous about recombination with outcrossing?

Tom: At the individual level? There isn’t any!


I wouldn't be so sure. It really needs careful consideration of what 'the individual' is here. Levels of selection are particularly important at this boundary, and mixing them up is a source of confusion. If one considers the haploid as 'the organism', and the diploids they form as binary pairings of such organisms, then the benefit of diploidy accrues to the partners jointly. We are much more used to thinking of diploids, or even massive somas made out of them such as us, as 'individuals' - entities with interests which are damaged by halving its genetic contribution to each offspring. But diploids really aren't individual, just persistent. From the 'haploid's eye view', their parting is just a return to the native state. In this context crossover, inasmuch as it helps even disjunction, is beneficial to the haploids because it allows this return. The slightly counter-intuitive feature of this is that the haploids returned are not the precise ones that went in. But there is no entity that really 'cares'. There are simply 4 ways to resolve a Holliday junction, and 2 of them result in crossover. The enzymes making the cut simply have 4 strands to hack at, and little upstream or downstream information. The alternative products are indistinguishable to anything that can do something about it.

Initially, at the individual level it was always only about DNA repair which only later represented an exaptation to what you already described as a proto-diploid state [which initially] could not be sustained indefinitely.

Personally, I think keeping a second chromosome set is too heavy a price to pay for its occasional use as a repair template. If you have a homologue for other reasons you can certainly use it for repair, but I doubt that repair is a sufficient force on its own to offset the cost. Prokaryotes certainly can't afford surplus DNA, and do fine without a second chromosome. If they are prone to badly-resolved DSB's, they die out rather than evolving a mid-cycle mechanism to deal with them. I don't see how an entire genome-doubling could be sustained in a primitive eukaryote for similar cost-benefit reasons.

But syngamising haploids come ready-provisioned. Much easier to sustain an extra chromosome if it comes packaged in extra cytoplasm. The haploids have garnered those resources separately.

Tom Mueller said...

Hi Allan

I am going to need to think about that.

If one considers the haploid as 'the organism', and the diploids they form as binary pairings of such organisms, then the benefit of diploidy accrues to the partners jointly.

I wonder if your subtlety of distinction is becoming semantic. This reminds me of one of my favorite professors who so many decades ago suggested that a chicken is nothing more than one egg’s way of producing more eggs. The problem here is that the identity of the haploid egg becomes diluted with each successive passage through a chicken. Actually I don’t really have a caveat with that observation. What we are talking about is a self-sustaining system. Who is to say the system begins and ends at the level of the individual egg much less the individual chicken. But at this point are we not talking about selection at higher hierarchical levels?

Personally, I think keeping a second chromosome set is too heavy a price to pay for its occasional use as a repair template.

That is exactly how I explain the transitory nature of the Spyrogea zygospore and even gametophyte dominance in "lower" plants. I need to wave my hands a lot when explaining the transition to the evolution of diploid dominance in plants. I suggest that at this point we are playing a molecular version of Jenga. At some point pulling out just one stick results in collapse, necessitating back-up copies. I know – I know… I did say I was waving my hands a lot. I just wanted my cherubs to remember the prevalence of gametophyte dominance at earlier evolutionary levels.

But syngamising haploids come ready-provisioned. Much easier to sustain an extra chromosome if it comes packaged in extra cytoplasm. The haploids have garnered those resources separately.

I need to think that one over. I think you may be on to something.

AllanMiller said...

Tom,

Me: If one considers the haploid as 'the organism', and the diploids they form as binary pairings of such organisms, then the benefit of diploidy accrues to the partners jointly.

Tom: I wonder if your subtlety of distinction is becoming semantic.


I don't think so, no. I think it changes expectations to look at sex from the haploid's viewpoint. If haploids came first, it cannot be wrong to regard it as a still cycling haploid-diploid-haploid-diploid succession. There is no point at which the cycle became diploid-haploid-diploid-haploid ... ! It may seem silly, but the perspective you take makes a difference, especially if you then decide that the entity of interest is halving rather than pairing. It's easy to be carried away with 'maximising genetic output' and regard diploids as being selected to take this to the extreme of 100% per offspring - ie asexual. It generates a mystery of sex, and that pesky twofold expectation, out of nowhere.

But for haploids - the simplest interaction would be a merger followed by immediate separation. As far as each haploid is concerned, the world looks the same - they still exist as 2 haploids, despite a transient binary phase. Now try a reciprocal swap during reduction. There are still two haploids, any gene on which is not particularly concerned with its upstream or downstream partners. Now try wrapping the germ line in a diploid soma, and using mitosis in the germ line to amplify the haploid genomes. Still, out of every Meiosis ll emerge those haploid gene sets, in precisely the same proportions as the first simple merger, if now confusingly scrambled. The world looks the same - haploids 'feel no force', like Einstein's falling roofer. They aren't selected to stay joined eternally. While the cycle continues, there is no point at which the haploid ceases to exist - it's just that in us, for 70 years or so they are locked in intimate binary association. The greater dominance of diploid phases over haploid in most organisms should not blind us to the fact that it's still haploids making diploids, and at no point (if my sequence is correct) did diploids suddenly start making haploids! It's just that mechanisms of final emergence have become more and more elaborate.

From the haploid perspective, there is not the same mystery to sex. It is a conundrum conjured by diploids who wonder why a diploid would bother splitting its genetic complement between offspring. But who cares about diploids? If they reduce to haploids, then those are genetic units with evolutionary interests; the diploids are simply very intimate associations of pairs of 'em, in some ways little different from lichenous symbiosis. One can treat the lichen as an organism for many purposes, but there's no point asking how it benefits the lichen to split into component fungus and alga at certain parts of the life cycle.

Of course, if a diploid line abandons reduction, it does become an evolutionary entity in its own right - but now, new dynamics take over. It can be hard to tell if this has happened, or whether reduction has simply been deferred.

Tom Mueller said...

Hi Allan

Your critique of a diplo-centric POV as a begging some sort of question seems, I think, to mirror what I was attempting to express (albeit more clumsily) by positing chickens as some transitory or ephemeral vehicle enabling eggs to produce more eggs.

So what is the unit of selection then? The cytoplasm? (appealing to metabolism-first proponents of abiogenesis) If we are going to flip-flop our terms of reference, why not re-consider genetic material as ancillary to a self-perpetuating cytoplasm? Or is that a bridge too far?

... or alternatively; the self-replicating system of serial alternation of egg-to-chicken? (appealing to some neo-Gouldian/Doolittlian hierarchy of selection)

I think there may be a few present that would jump all over that last one.

You serious? You think we could glean a paper from all this?

Tom Mueller said...

oops - hit send too soon

I meant ... or alternatively; the self-replicating system of eggs to eggs via transitory chickens? (appealing to some neo-Gouldian/Doolittlian hierarchy of selection)

Actually, this alternative is not really "alternative" to the first suggestion of cytoplasm as a unit of selection.

If I understand you correctly, the serial dilution of genes via sexual recombination is misunderstood to be some sort of "cost" as opposed to really a "means" of keeping the entire self-replicating system running.

Again, it just depends on your POV. Did I understand you correctly?

AllanMiller said...

Tom, re:So what is the unit of selection then? The cytoplasm?

Absolutely not! This is a version of gene-centrism.

I think the principal unit of selection is the haploid genome, in the matter of sex. When they form a diploid, this becomes a unit too, but ultimate release for the haploids restores their status. In a simple system, the evolutionary dynamic consists in tuning haploids for both their individuality and their co-operativeness, because they have a 2-phase life cycle, with selection in both. One phase has come to domiinate - the diploid, usually.

If a pair chain themselves eternally in an asexual diploid, the haploids essentially disappear (with numerous interesting consequences). If, on the other hand, they reciprocally recombine, units of selection smaller than the haploid genome are uncovered - Dawkins's selfish gene is released. Selection actually involves multiple levels, of course.

One surprise to me is that Dawkins's reference points - Williams, Maynard Smith, Hamilton, architects of the 'gene-centric' view - did not appear to reach this obvious conclusion from their own work, all of them having much to say about sex. They seem to me to apply an incorrect version of their own gene-centric view, which bumps up against diplocentricity – taking a supposed gene’s eye view, why would a gene suffer a 50% reduction in its chance to get into the next generation? This is Williams’s Cost of Meiosis, but despite being supposedly expressed at gene level, it is a disguised ‘diplocentric’ argument. Haploids which merge then separate are not suffering a 50% reduction in anything. Haploids are Williams’s fundamental evolutionary alleles, not anything smaller or larger, although both smaller and larger units are involved in the dynamic as the system progresses.

AllanMiller said...

If I understand you correctly, the serial dilution of genes via sexual recombination is misunderstood to be some sort of "cost" as opposed to really a "means" of keeping the entire self-replicating system running.

Not quite. It’s more that the costs are illusory.

The Twofold Cost is expressed in two ways: as a cost of males, and as a genetic cost of dilution (Cost of Meiosis). They are often felt to be different ways of saying the same thing, but they aren’t. One is a consequence of there being two genders, the other of there being two genomes in a diploid.

The Cost of Males is actually not a genetic cost at all – after all, the genetic complement of the two sexes is pretty much equal. They differ only in the ‘costs’ of production. Since genes invest equally in both the genders, spending 50% of their time in each, they invest in two strategies, so males aren’t costly to genes.

However, an asexual ‘female’ can produce only female offspring, and so can double her output every generation over one producing males. But in order to make this switch, she must leave the population. Becoming an asexual female means that no genes can flow from you – you are not the same species, even if you are ecologically indistinguishable. The Cost of Males is actually The Cost Of Not Becoming A Different Species, which is just bizarre!

The Cost of Meiosis is different. While full meiosis exists, complete with independent segregation and crossover, each genetic locus so exposed has notional ‘interests’, if one is not too hostile to Dawkins’s way of looking at things. Each locus has ‘levers’ with which it can push itself into the next generation. It can make the organism the right shade of green, more disease resistant, or whatever. It can even take a potshot at its homologous sequence and distort its Mendelian transmission. But it is a mistake to equate this last with a summed ‘desire’ of all genes to abandon meiosis altogether. Because in abandoning meiosis, you abandon the very thing that makes you a locus with interests! It’s as if a gene reaches out for the lever, grabs it, and … boom! It ceases to exist. It has been subsumed into the Borg. It ... Evolutionary alleles don’t keep on being evolutionary alleles when the mechanism that makes them so ceases to operate.

Genes for being asexual don’t spread around the population. They create separate populations.

Tom Mueller said...

Allan, I confess I was skeptical at first - but I think you may be on to something.

You intend to publish any of this?

Larry Moran said...

@Tom

Allan is just summarizing the work of Maynard Smith and others from many decades ago.

http://sandwalk.blogspot.ca/2014/03/everything-you-thought-you-knew-about.html

For any given organism in a sexual population, it appears to be beneficial to reproduce asexually because it avoids all the hassle of finding a partner and fusing your genomes. Thus, one successful yeast α cell could continue to pass on its genes forever, and many do.

It's not clear what advantage is gained by fusing with an "a" cell and undergoing reduction division but it IS clear that there is an (approximately) twofold cost to doing this. That's because if each of the cells divided twice there would be eight offspring instead of only four from sex.

Allan is commenting on the confusion over this cost. Far too many people equate it with very complex organisms where the differences between the sexes are great and one of them bears a heavy burden of sexual reproduction. That's pretty much irrelevant when we're discussing the origin of sex. (Although it does raise a question about the usefulness of men.)

AllanMiller said...

Larry, re:
It's not clear what advantage is gained by fusing with an "a" cell and undergoing reduction division but it IS clear that there is an (approximately) twofold cost to doing this. That's because if each of the cells divided twice there would be eight offspring instead of only four from sex.

I think this misses my central point. I’m not quite summarising Maynard Smith, I’m arguing that both his and Williams’s costs are illusory from the perspective of the haploid. The costs are an artefact of a ‘diplocentric’ point of view. If 2 free-living haploids merge then reduce, it would be illogical to argue that the transient diploid so formed suffers a twofold cost from the reduction. As the cycle elaborates, and diploids become persistent entities in their own right complete with mitoses in the diploid line and cytoplasmic asymmetry, this relationship does not change. Diploids don’t suddenly become the governors of the sexual transaction, at any point. There is no component of a diploid genome which does not ‘belong’ to one of the constituent haploids. Diploids are binary organisms.

Say 2 haploids fuse and the diploid so formed undergoes 1 mitosis, and one of these diploids has a mutation that prevents reduction. Allow 2 more mitoses in each line. We now have 8 diploid cells. 4 of those reduce to give 8 haploids. The others are stuck as diploids forever. Now the 4 ‘asexual’ cells have exactly the same number of chromosomes, pretty much the same amount of cytoplasm, etc, as those 8 haploids. So it would appear that the material costs are the same. And, indeed, there are more cells. It’s just that we are used to regarding the diploid as the central actor, and their purpose as to produce more diploids, which can obviously be done by mitosis alone.

In the reducing lineage, the original 2 haploids have gained 4 copies each. The fact that they did so in tandem does not give the tandem unit interests over and above theirs.

AllanMiller said...

I’d add that I don’t see this as a trivial ‘either-or’ POV argument. It appears to me to be a fundamental fact that diploids cannot evolve sex. Leaving aside the question of how we got such a diploid in the first place, and any twofold notions, any diploid that reduces will find a world devoid of haploids with which it can fuse, other than its erstwhile partner. This applies to long-term asexual descendants of sexual ancestors as well as endomitotic diploids. The sexual world can only be built from the haploid up. Diploidy can be an adaptation for haploids; the reverse cannot be true.