Nelson: Developmental Systems Drift

Conversation
41

Maybe this will help.

In several of his publications (roughly, from 1998 until his death in 2012), Carl Woese argued that LUCA never existed, as a discrete cell ancestral to all extant life. “The universal ancestor,” he wrote, “is not a discrete entity.” https://www.pnas.org/content/pnas/95/12/6854.full.pdf If a cell is anything, it is a discrete entity. But Woese jettisoned LUCA. His alternative notion was a community of progenotes, with no single progenitor entity. The community’s origins disappear into abiotic processes, and may have occurred in multiple or independent spatiotemporal sites.

For Woese, a key piece of evidence for this scenario were the structural differences among ribosomes in the three domains:

The ribosomes of the three domains “crystallized” independently of each other. Thus, all extant cells having a ribosome – for Jonathan Eisen (see his slide, above), a marker of universal monophyly, or of LUCA’s real existence – was not, for Woese, evidence of LUCA’s real existence.

Every object at the threshold between the abiotic and biotic realms comes with two associated probabilities: (1) the probability of its origin abiotic realm-up from physical and chemical, but NOT biological, processes (because biology doesn’t exist yet) and (2) the biotic-down or phylogenetic inference probability. Eisen says “NO WAY” that multiple independent origins of ribosomes could have happened, but that inference rests on (2), the biotic-down phylogenetic inference probability, not (1).

Problem is, if you look at the threshold from the perspective of deterministic chemistry, and raise the (1) abiotic-up probabilities towards unity, as for instance Christian de Duve wanted to do, that necessarily erodes the (2) biotic-down probabilities. If ribosomes are bound to happen, once we figure out the abiotic pathways (the physical recipe) leading to them from chemistry, they cease to be reliable phylogenetic markers of LUCA. “Here,” says chemistry, “have a ribosome on me. You’ll need it to be a cell on this planet.” Everyone gets a ribosome, thanks to this (currently unknown) abiotic pathway.

SO – defining universally-distributed molecular characters, such as ribosomes, as necessarily stemming from LUCA wipes out even the possibility of a Woese-type scenario. Evidence from extant cells that they might not descend from LUCA disappears, not because we actually know that they descended from LUCA, but because we have stipulated that LUCA existed, ergo… Anyway, putting data inside what I call “the LUCA horizon” changes logically how one thinks about it.

Art, none of this has anything to do with DSD. I was miffed because Steve Matheson said I was misinforming students. My conception of neo-Darwinian theory includes thinking about how terms such as “homology,” critical to understanding that theory, are defined. In a classroom setting I have a chance to explain what I mean, and to forestall misunderstandings. So I’ll think seriously about changing the offending slide, for contexts like this where I can’t be there to say ‘this is what I mean by neo-Darwinian.’

Chris, thanks for the assistance, but I’m talking strictly about naturalistic evolutionary theory. Allow for design and everything changes.

42

I reject all the claims. We think ribosomes have a single origin because of their structural similarities, particularly similar sets of rRNAs. This has nothing to do with assumption of a LUCA, but it does support the inference of a LUCA. Nor is there any reason to suppose that ribosomes arose abiotically, especially if you consider the RNA world hypothesis.

Out of curiosity, what is your preferred definition of “homology”?

2 Likes
43

I like Van Valen’s definition of “homology” – “correspondence due to continuity of information.”

But really? Homology is a concept which, if you think about it long enough, will drive you crazy:

http://www.arn.org/docs/odesign/od182/hobi182.htm

44

Interesting paper, but you have forgotten to explain what it has to do with ribosomes. Doubtless an oversight, and easily fixed.

In what way does that differ from “similarity due to common descent” other than in the vagaries of wording?

Are you claiming that you and Jonathan Wells are crazy? Did you intend some additional point by linking the article?

2 Likes
45

There are huge differences between hammerhead ribozymes and the peptidyl transferase. Whatever point you want to make with this example, @pnelson, it doesn’t help your case here.

2 Likes
46

The Hammerhead ribozyme isn’t a ribosome.

Arguing that some ribozymes have multiple origins doesn’t mean that ribosomes have multiple origins, just like some people having dual nationalities doesn’t mean that @pnelson has dual nationality.

3 Likes
47

To my knowledge increasingly ancestral nodes in independently derived trees of ribosomal RNAs for different clades, nevertheless exhibit ancestral convergence. If that isn’t evidence for common descent, it’s hard to know what is.

2 Likes
48

Of course.

But the point of linking to that article is to encourage critical reflection. How do we know that the peptidyl transferase center (PTC) had a single origin? – the standard view, which holds that the PTC is a universal homology, in the sense of unifying all terrestrial biology historically. Your PTC, and E. coli’s PTC, prove that you and E. coli are kin, if we go far enough back.

Answer: their relative degrees of complexity. The hammerhead motif is small. The PTC is big (thousands of nucleotides long, complex architecture).

Okay, thought experiment: jump into my time machine with me, and travel to 2095. High school AP biology class, lab section. Today’s exercise: generate the PTC from activated mononucleotides, following the protocol in the lab manual. By the end of the day, students should have a functioning PTC in their beakers. Just follow the protocol, which specifies the deterministic pathway from the PTC’s nucleotide constituents to its functional architecture.

Now the reason this is science fiction is (in 2020) we don’t actually know the pathway from nucleotides to the PTC. So when one reads the abiogenesis or RNA World literature today, it’s luck. A one-off event, or what Israeli philosopher of science Iris Fry calls “a near-miracle.” Some damn thing happened, the PTC arose via an unknown chemical pathway, just ONCE (nature hit the lottery), and here we are.

But the whole point of doing research on the origin of life (OOL) is to reduce the “luck” factor, or, more formally, to limit appeals to unrepeatable, untestable contingencies. What happens when luck goes down and knowledge goes up? Things can happen more than once. Indeed, the “same” event will happen more than once. Christian de Duve argued vigorously that the road to understanding the OOL lay in doing whatever was possible to raise the probabilities of the individual events, from chemistry to cells, so that near-miracles were not the content of abiogenesis theories.

SO – my point is (again) to encourage thinking about the conceptual connections between phylogenetic inference and theories about the OOL. If there is an unknown deterministic pathway to the PTC, our current inferences that the PTC is a universal homology will erode. Bound to happen.

There is a deterministic recipe for the hammerhead ribozyme, which tells you not to use that structure as a phylogenetically informative molecular character. Apply the same analysis to the PTC, and ask yourself what you actually know about how the PTC first appeared in Earth history.

49

Leigh told me, with a twinkle in his eye, that he wanted a definition of homology which would include Agassiz, Owen, and Cuvier. “Continuity of information,” he explained, could mean continuity in the mind of God. Leigh (as you know, John) was a rebel who marched to his own drum.

50

Yeah this whole thing assumes the ribosome is an outcome of some sort of soup pot-chemistry (that is the insinuation of the phrase “from activated mononucleotides”) as opposed to an outcome of evolution operating on a population of (perhaps cell-like) entities that didn’t have translation.

The RNA world hypothesis is not really postulated to be a hypothesis for the origin of the translation system by hydrogencyanide chemistry. It’s supposed to be a hypothesis for the origin of an RNA-based genetic system(and similar such scenarios are usually called “genetics first”, not “ribosomal translation first”), which could substantially predate coded protein biosynthesis.

The point of abiogenesis research is to get to evolving life(which is where some RNA world proponents posit a kind of protocell-stage where fatty acid vesicles would encapsulate simple ribozymes), which they think is something that is recognizably alive and able to evolve. From such a stage and onwards, explaining the kinds of living cells that now populate the biosphere(and by inference the universal common ancestor) becomes a job for evolutionary biology. Not “activated mononucleotide” soup-chemistry.

2 Likes
51

Yes, but it’s a thought experiment. Make the necessary amendments to the details. The RNA World hypothesis, and abiogenesis theories generally speaking, aim to provide the highest probability pathway from chemistry to cells.

52

My critical reflection first led me to wonder if you knew that hammerhead ribozymes aren’t ribosomes.

I think that’s a wrong answer.

PTC is considered to have a single origin because phylogenetic analysis indicates this (eg here). Similar analysis of hammerhead ribozymes suggests the opposite (eg here). It’s not the degree of complexity, it’s the ability or inability to create a phylogenetic tree.

1 Like
53

More on the same point (i.e., complexity as a diagnostic of unique occurrences):

The specific content of the genetic code and probably the unique role of DNA in cellular organisms are contingent states of affairs that surely must be understood in terms of history, in the sense that these outcomes are not simply the necessary consequences of evolutionary laws applied to some class of carbon-rich planets that maintain, for a certain extended interval, a certain amount of surface water and a fairly temperate regime. (Burian 1988, 264)

Colin Patterson of the Natural History Museum (London) wrote extensively about the foundations of the homology concept. From his treatment (1988, 615) of molecular homology:

Convergence between molecular sequences is too improbable to occur, just as similarity between sequences is too improbable to be explained, except by common ancestry…This is the argument from complexity: if two structures are complex enough and similar in detail, probability dictates that they must be homologous rather than convergent.

Here is Jonathan Eisen’s UC-Davis lecture slide on the improbability of the ribosome. Note that he does not say ‘we know the PTC and the ribosome built around it point to universal common ancestry because we can build phylogenetic trees from rRNA sequences [16S, 23S, etc].’ Rather, his argument is this: you will NEVER get this molecular machine evolving more than once on Earth.

Screen Shot 2020-05-05 at 10.26.38 AM
Screen Shot 2020-05-05 at 10.26.38 AM1816×1378 276 KB

Burian, R. 1988. Challenges to the evolutionary synthesis. Evolutionary Biology 23:247-269.

Patterson, C. 1988. Homology in classical and molecular biology. Molecular Biology and Evolution 5:603-625.

54

One can construct phylogenetic trees with sequence data which we know did not evolve from a common source:

“Phylogenetic reconstruction using the complete genome sequence [of φX 174] not only failed to recover the correct evolutionary history because of these convergent changes, but the true history was rejected as being a significantly inferior fit to the data.”

From here:

Look at the alignment in the hammerhead paper. If one fed those sequences into any phylogenetics tree-building software, it would kick out a common ancestry pattern, even though Salehi and Szostak “bred” the hammerheads as independent populations.

55

Sure, but then you don’t have a basis for arguing that abiogenesis is some sort of central pillar in evolution, just because life shares universal common ancestry. You’re the one making the assumption that supposedly connects the two, it doesn’t follow from the theory of evolution even including common ancestry. You are the one including premises in that argument which are not intrinsic to evolution by common descent.

2 Likes
56

One can, of course. But should we draw any major lessons from that? Bull et al. think not:

57

Anyone else interested in developmental systems drift? What it means from a design perspective? Or any other angle?

58

How prevalent is it compared to developmental system non-drift? I think of Pax-6, for example.

1 Like
59

Great question, it’s exactly what I would explore in a real conversation about DSD. Pax6 and other remarkable examples of deep homology are in some sense the very opposite of DSD and yet both seem present in phylogeny.

60

Consider also the conservation of Hox expression in linear sequence both in the genome and the embryo.

1 Like