This is a good question with a long technical answer, but it can technically go as low as 0%, because two proteins with 0 sequence similarity can adopt basically identical structures (and two proteins with 20% or more sequence similarity can adopt very different structures). But the variations in structure between proteins with essentially no sequence-homology can still exhibit significant phylogenetic information, and thus imply relatedness even with very different primary sequence.
There is a sub-branch of the field of phylogenetic inference dealing with determining homologous relationships from protein 3-dimensional structures. Combined with analyzing how the proteins accomplish their functions at a biochemical level, scoring their similarities by various structural measures, (including perhaps with things like gene synteny or interaction partners) and considering their distribution in different genealogically related species, a phylogenetic relationship between proteins with completely different amino acid sequences can still be established. No simple or easy answers here Iām sad to say.
Thatās extremely interesting and Iām grateful for the explanation. Are structural homologies considered presumptive evidence for common ancestry in the same way that sequence homologies night be?
It very much depends on a lot of details to distinguish between common ancestry and convergence. Those are two different hypothesis that can be texted with data.
I just finished reading the paper. My first question is an overarching one: why is it simply assumed that one homologous protein can evolve into its homologue? Why does nobody feel the need to detail an evolutionary pathway before declaring victory?
Well, thatās a lot less helpful and informative than I would have hoped.
The paper lists a number of homologies between flagellar proteins and other proteins that perform other tasks. It also lists homologies between proteins used in the flagellum itself. At no point do they explain how these proteins supposedly evolved into their flagellar homologues. Itās simply assumed that they did, or at least could. Why is that assumption justified?
A commendable attitude I have to say. Thereās absolutely nothing wrong with asking questions.
Yes this is a topic that sows an awful lot of confusion, among other things because scientists honestly arenāt being all that clear on this topic either.
I think you need to separate out these two questions:
What is the selective effect of most mutations that occur and go to fixation across some organismās genome, over generations?
How does some specific, clearly adaptively beneficial organismal attribute, originate and subsequently evolve?
The idea is that a majority of mutations that go to fixation are selectively neutral, but there is still some small but significant portion that have selectively beneficial effects, and thus can contribute to optimizing some adaptation.
But youāre probably also thinking about two other things, which are rarely well articulated in these matters. You want to understand the relationship between evolutionary innovation, molecular complexity, and adaptation.
To what extend do mutations that occur and go to fixation constitute innovative mutations?(how often do mutations result in novel biological functions?).
To what extend do mutations that occur and go to fixation contribute to increases in molecular complexity? (like adding more functional genes to the genome, adding more proteins to some existing structure, and ultimately result in more complex multicellular organisms with multiple distinct organs and all that stuff)
To what extend do mutations that occur and go to fixation contribute to increases in reproductive success?
Itās important to understand that these three issues are not the same thing, and the relationship between them is complicated, depends on circumstance, and can be found anywhere from being correlated to anti-correlated.
Complexity can go up while fitness goes down, while number of total functions remains the same. Think of mere duplication resulting in multiple unnecessary gene-copies that negatively affects some organisms metabolic budget by the cost of expressing them, leading to a slight fitness decline. Genomic complexity has gone up as the number of functional genes and genome size has increased, but it has incurred a slight fitness loss. No new function was gained.
Complexity and fitness can remain the same while number of total functions goes up.
Think of point mutations that make some enzyme able to act on a novel substrate and break it down without altering itās existing function, but the organism has no use for these novel break down products. Number of functions have gone up, but the organism is unaltered in terms of fitness or complexity.
Fitness and functions can go up while complexity decreases.
An organism is suffering deletions in some gene, which makes the gene relocate to another part of the cell when expressed, which alters itās morphology so it becomes better able to resist an antibiotic. Itās evolved a new function through a deletion(decreasing itās genomic complexity), and this function happened to be beneficial.
And of course many other variations on those themes.
One can imagine, and find innumerable examples of mutations that have such effects. There IS NO easy or obvious relationship between fitness, innovation, or complexity. They can all go up or down independently of each other. Sad but true.
Evolution is not thought by any extant evolutionary biologist to constitute one long unobstructed gain in organismal complexity (or reproductive fitness) through the history of life. Lots of existing genes can be duplicated and degrade to mutations, and this can be beneficial, or it can be deleterious, or it can be neutral, and new functions can be found once in a while that might suddenly become beneficial and be super-optimized by positive selection.
That said, neutral processes can contribute to the evolution of both complexity and novel functions in something called constructive neutral evolution. Selection still plays a role in this process, but itās mostly through so-called negative selection. Removing deleterious variants while merely retaining still functional ones.
Itās important to understand that complexity is not necessarily beneficial, nor necessarily deleterious. It is highly context-specific. Complexity can result in adaptations, but it can also be mal-adaptive. There is no simple relationship between fitness and complexity.
Iāve posted this figure before that is supposed to explain how even ādevolutionā (not a real term in biology, just borrowing it from Behe) can result in gains in functions and increases in genomic complexity, while being almost entirely driven by a combination of neutral changes and negative selection:
Squares represent genes, colors and intensity represent functions and their degrees(brighter color = higher degree of function). Red rectangles highlight what is being duplicated and passed on.
This is āadaptive devolutionā of increased complexity, and new functions, by mostly ādegradingā and mostly ābreakingā genes. Because these extra genes are costly to express, their death is adaptive, and so is the eventual deletion of them. But because the still functional copies continue to accumulate deleterious mutations, as these are are more frequent than beneficial ones, their duplication is also some times adaptive(more expressed genes compensates for each individual gene being weaker).
Eventually over many generations a previously dead gene locus, a black square (effectively having become non-coding DNA) evolves into a de novo protein coding gene (purple square, Function B). This new functional gene suddenly comes under strong positive selection so is quickly improved over subsequent generations. So one new function is evolved and enhanced, while all the rest degrades and breaks. The net result is more complexity and more functions than there was to begin with. And it happened almost exclusively through neutral and adaptive degeneration. There was one innovative mutation among thousands of degenerative ones.
There are two possible answers to this, and of course both depend on what exactly you think would qualify as a complex system and what you think counts as an explanation.
Most of the evidence for evolution is inference from comparisons of different organisms. Scientists infer that certain things evolved (and how they evolved) from being able to derive the relationships between the organisms that carry these attributes using phylogenetic methods.
That said, there are of course experiments and observations of ācomplex systemsā that evolved. But this is where much depends on what you mean by a complex system. If you think we need to show by observation the evolution of an entire bacterial flagellum from some state where none of itās constitutent proteins even exist (or where none of them have come together), then that canāt be done as that simply takes too long. And thatās really not an excuse, because we really do have to accept that some large-scale developments just take much longer than can be directly observed.
Iāve never seen the formation of an entire mountain range, I havenāt seen the emergence of an entire forest, or a river cut a path through hundreds of meters of rock. But we can observe smaller-scale changes in the present, and with those infer that they can add up to larger-scale changes in the future if they are allowed to continue at a similar rate.
There are ways of elucidating that certain things occur on long timescales, and have occurred in the past, that allow us to conclude this with good confidence, without us having to see it happen in real time. There are such methods used in comparative genetics that allow scientists to infer not only THAT certain structures evolved, but some times even how.
Thereās a method called ancestral sequence reconstruction, where scientists can test evolutionary inferences using phylogenetic methods. They can literally use phylogenetic trees to derive what ancestor states would have looked like, infer their functions, recreate them in the laboratory and test them to see if their inferences are right (and test to see if the ancestor state really would have been functional). They can then test the functional and selective effects of historical mutations that occurred between the reconstructed ancestor and itās extant descendants.
For example, scientists from the Thornton laboratory have recreated small pieces of the evolutionary history of the vacuolar ATP-synthase molecular machine, which you can read about here:
They have many other good publications on ancestor reconstruction where they show different new functions and proteins evolved over time.
I think the statement that there are ārelatively fewā sequences that result in viable proteins is vague. And I think thereās good evidence that it is nowhere near so low that new proteins canāt evolve. Iād be happy to discuss more of that if you are interested.
Generally speaking the case for homology (the inference that the proteins in question really are homologous, that is they share common ancestry and so evolved from a common ancestor) is based on being able to show that there is significant nesting hierarchical structure in their shared similar characteristics (for proteins this can be derived from analyzing and comparing their amino acid sequences and/or the DNA sequences that encode them). Most of these methods involve being able to derive a phylogenetic tree from the sequences of the proteins and seeing how well-supported the tree is by various statistical measures.
The case for actual evolutionary relatedness then can be further supported by showing that the tree implied by one protein, is largely reflected in the trees derived from other proteins. This yields a concept known as consilience of independent phylogenies. The question becomes, why should the phylogenetic tree derived (using some systematic algorithm) from the sequences of one gene, yield a tree with a similar branching topology to an independently inferred phylogenetic tree derived from an entirely independent gene?
This kind of comparison can be done both within and between genes. When similar trees are consistently recovered from different parts of the data, thereās really no other good explanation for this fact than the fact that these different parts of the data all really did go through the same genealogical process of branching descent with modification.
Saying that anyone who disagrees with you isnāt a Christian is a real conversation-stopper. Do you really want to do that?
More briefly than possible, Iād say. You just canāt go around saying that a Christian must be an IDer or creationist (not sure which you intended).
Universal common ancestry, I believe. How he thinks ID works within that framework has never been clearly stated.
Some lens crystallins in various taxa (birds among them) would be another example. Some of them also function (presumably their original functions) as enzymes elsewhere in the body.
Daniel, you are making good points. I am also a ālaymanā, and I would say these guys have a very difficult time, bringing the knowledge down from the top shelf (itās like Computer Geeks talking in ācodeā language!) Simple answers are are good place to start, and usually better in all areas of life. That been said, there is value in trying to pick up what is being said. I also agree, totally, that debates are extremely valuable. In Proverbs it saysā¦ āthe First to present in Court sounds Right, then the cross-examination beginsā. @swamidass Josh has the guts to at least engage, respectfully, and that is very Rare. I donāt know if you are relying on your Intuition, but if you are I would Encourage you in that endevor. I would say it is Clear, that all/most of these āthingsā were Designed (and Evolution had nothing to do with it), you will be Fully Challenged here on your Intuition/CommonSence/ClearlySeen opinion. The Behe/Swamidass 2nd, 90 min video, pushed me even further to the Design side of the spectrum!
Iām not talking about validating evolution, Iām talking about donstrating that it is even possible on a large scale. Without a detailed, stepwise account for a particular case, how can you call it anything more than speculation?
What does that even mean? How many steps for you would qualify as stepwise? Iāve seen plenty of what I would call stepwise explanations. But then a skeptic, would find one area where we arenāt exactly sure and go ahah!
@Daniel_Arant it seems the same could be said of design, right? Without a detailed stepwise account of design, how do you have anything more than speculation? The fact of the matter is that we have far more details on how evolution has taken place (even though we do not have the whole story) than we have of Godās design.
If this were a conversation about religion and christianity, then it would be a conversation stopper. But thatās not what this conversation is about. You seem determined to be disagreeable?
I appreciate the explanation. Forgive me if Iām oversimplifying, but it sounds like yes, evolution is now viewed as a mostly random process. Natural selection plays a relatively minor role. Does this not increase the difficulty of discovering new functionality (which I take to mean everything from a new ezymatic function all the way to a new body plan?)
Iām also struggling to understand how so many neutral mutations can become fixed without any selective pressure. I think I understand the concept of genetic drift broadly, but it must include an awful lot of scenarios in order to be able to explain so much genetic fixation, no? Are there any observations of the rapidity with which neutral mutations become fixed?
Think about it this way. Letās say that to āwinā you have to be able to sink 100 free throw baskets (i.e. get 100 specific mutations). If you have 100 throws, how difficult is this? Pretty hard. Even professionals might struggle.
Now, what if you have a 100,000 throws to get 100 baskets? Well, now it is much much easier. The vast majority of mutations can be off in random directions, and it doesnāt really matter. Though the number of random throws is dramatically increased, the chance of getting the mutations you need has increased a great deal.
So yes, most mutations are not useful, and are not selected. But that also means that only a few mutations are required for important evolutionary changes, and there is a lot of extra shots to find those mutations.