|
Register | Sign In |
|
QuickSearch
EvC Forum active members: 65 (9164 total) |
| |
ChatGPT | |
Total: 916,903 Year: 4,160/9,624 Month: 1,031/974 Week: 358/286 Day: 1/13 Hour: 1/0 |
Thread ▼ Details |
|
Thread Info
|
|
|
Author | Topic: Random mutations shot down on this site. | |||||||||||||||||||
Percy Member Posts: 22504 From: New Hampshire Joined: Member Rating: 4.9 |
Hoot Mon writes: Now, on the issue of speciation and the role of drift, I wrote in Message 15:
quote:They are taling about speciation here. The full text of the article is only available to subscribers, and the abstract doesn't even mention genetic drift. Could you provide relevant excerpts from the article that support your point?
I wish I could get you to examine Smith and Eyre-Walker’s Science article, reference above. They found 55% of the protein differences between two fruit fly species to be attributable to drift. And please see the work of Daniel Hartl, also referenced above. He often stresses the non-adaptive aspects of speciation. Right. We understand the point you're making, we're just trying to determine if the article actually supports your point. Genetic drift by its very nature would seem an odd candidate for the cause of much speciation. --Percy
|
|||||||||||||||||||
Percy Member Posts: 22504 From: New Hampshire Joined: Member Rating: 4.9 |
Doddy writes: I'm with Hoot Mon on this one. I'm not so sure about that, but first let's revisit your erroneous statement that Crash tried to correct:
For there to be random mating, each tree in the forest should have equal chance of being pollinated by every other tree, which does not happen due to the long distances in the population. Can you imagine any species in the known universe where an individual has an equal chance of mating with all other living individuals of its species? No, of course not. So obviously that's not what random mating means.
This is not sexual selection, because there is no selection of mates by either party. Just luck of the wind and insects usually. I'm not sure there's wide acceptance of such a thing as sexual selection in plants, though this has certainly been proposed since there is sexual dimorphism and other sexual differences in some plants. But I think in general people would have to agree that sexual selection is not widespread or easy to detect in the plant kingdom. But you're using the example of plants in support of an argument that Hoot Mon advanced concerning animals. In essence Hoot Mon is saying that sexual selection is not a type of natural selection, which is nonsense. --Percy
|
|||||||||||||||||||
Percy Member Posts: 22504 From: New Hampshire Joined: Member Rating: 4.9 |
Doddy writes: Let me quote the definition on wikipedia:
quote: That is what random mating means. And, no species does it...so what? I said that no species does what you described as random mating, not what Wikipedia describes as random mating. Your description of random mating is not the same as Wikipedia's. Your description of random mating was incorrect. This is what Crash and I have been trying to tell you. Here's your original description of random mating from Message 35:
For there to be random mating, each tree in the forest should have equal chance of being pollinated by every other tree, which does not happen due to the long distances in the population. That's not the definition of random mating. Obviously there is no population of anything anywhere where all organisms are equidistant from all other organisms of the same species. Such a definition of random mating is nonsensical and useless. In other words, it is self-evidently wrong. --Percy Edited by Percy, : Grammar.
|
|||||||||||||||||||
Percy Member Posts: 22504 From: New Hampshire Joined: Member Rating: 4.9 |
Doddy writes: I must be missing something.... Yes, the obvious. Random mating does not mean equal opportunity to mate with all organisms in the population regardless of geographical distribution. Definitions of random mating do not usually mention the incredibly obvious effects of geographical distribution because they are so, uh, incredibly obvious. There's a question I've put to you several times, and I think that if you try to answer it that it will help you see where you're misunderstanding things. Name a species for which the opportunity for an individual to mate with all other individuals of the same species is equal. You won't be able to do it. No such species exists. Now ask yourself why random mating is described as if it were something that is actually observed? The answer is that it is observed. Random mating exists and has been observed. So since by your understanding of the definition of random mating it isn't possible and so shouldn't be observed, your understanding must be incorrect. Leaving out pieces of the definitions you gave doesn't help your cause. For example, this is the full text of your last definition: "In the ideal case, each individual in the population has the same probability of mating with every other individual. In practice, no selection influences the matings that occur." The Wikipedia definition is better (Panmixia - Wikipedia), as is the definition at this very site (http:///WebPages/Glossary.html#R). --Percy
|
|||||||||||||||||||
Percy Member Posts: 22504 From: New Hampshire Joined: Member Rating: 4.9 |
Hoot Mon writes: I think you are wrong. I bothered to go get the whole article through my public library. Try it. You would be impressed with their drift v. selection model. Obviously Wounded King already has a copy of the article, for he provides far more information from it than appears in the abstract. Could you please provide excerpts of the sections of the article that you think support your position? --Percy
|
|||||||||||||||||||
Percy Member Posts: 22504 From: New Hampshire Joined: Member Rating: 4.9 |
Doddy writes: Percy writes: Random mating does not mean equal opportunity to mate with all organisms in the population.. Sorry, I'm fairly sure that it does. All genetics simulations that I've encountered do this when set to random mating. Any simulation that does this is only making an approximation of the real world. And it can be especially helpful to ignore considerations of geographical distribution when that is not a factor under study.
As one of the assumptions of the HWP, random mating is described. Infinite population size can't be observed in any species either, but that doesn't mean it is a useless definition. ... But, it appears we are both using different definitions - I refer to the theoretical, and you to the real-world. Ah, okay, I see what has happened. You're interpreting what I'm saying in terms of idealized models, while I'm talking about the real world. But it was clear from the beginning that I was talking about the real world, because so were you. You were talking about trees in a forest in your Message 35. I quoted what you said about trees just above my response in my Message 40, and here it is again:
Percy writes: Doddy writes: For there to be random mating, each tree in the forest should have equal chance of being pollinated by every other tree, which does not happen due to the long distances in the population. Can you imagine any species in the known universe where an individual has an equal chance of mating with all other living individuals of its species? No, of course not. So obviously that's not what random mating means. I don't know where you got the idea that my comments had somehow shifted focus from your concrete example of trees in a forest to the other discussion concerning idealized models, but no such shift ever occurred. --Percy
|
|||||||||||||||||||
Percy Member Posts: 22504 From: New Hampshire Joined: Member Rating: 4.9 |
Hi Hoot Mon,
Let's compare what you claimed way back in Message 15:
Hoot Mon writes: Two British evolutionary biologists, Nick Smith & Adam Eyre-Walker (Adaptive protein evolution in Drosophila, Nature, 2002) have studied adaptive protein evolution in two fruit-fly species D. melanogaster and D. yakuba. They determined that the two species separated about 6 million years ago, and that 45% of the separation could be attributed to natural selection. The remaining 55% of the protein differences were determined to be selection-neutral, and these difference were attributed to genetic drift. And here's what you just said in your most recent reply in Message 59:
Hoot Mon writes: To me, this is a credible attempt to differentiate, as casual factors attending microevolution, the actions of natural selection from those of drift. If I'm interpreting this correctly, you're portraying yourself as never having misconsrued anything in the paper, and you won't do it again. --Percy
|
|||||||||||||||||||
Percy Member Posts: 22504 From: New Hampshire Joined: Member Rating: 4.9 |
Hoot Mon writes: But you obviously didn't read my Message 43, in which I made clear to Quetzal my mistaken conflation of speciation with microevolution. You're just ankle biting again. If he's ankle biting then so am I, but I'm not and he's not. What you're offering in your last couple posts is just straightforward dissembling. First, just look at your claim about conceding to Quetzal in Message 43. Your very next message, Message 44, is to Wounded King telling him he's wrong and to go bother getting the article, with no substantiation and as if you hadn't read a thing he'd posted where it was eminently clear he had the article already in hand. Second, you wrote a whole post, Message 59, that was mere revisionist history, arguing as if it were Wounded King who had misinterpreted the article. The qualities of arrogance and chutzpah are not pretty and certainly not worthy of anyone capable of tackling the complex topics we address here at EvC Forum. I suggest just focusing on dispassionately discussing the issues and letting the readers of these threads make up their own minds. --Percy
|
|||||||||||||||||||
Percy Member Posts: 22504 From: New Hampshire Joined: Member Rating: 4.9 |
Hi Hoot Mon,
So now that the point you raised concerning that paper in Message 15 has been addressed, do you have any further points to raise concerning the characterization of RandomMutation.com is for sale | HugeDomains provided by the many respondents to Message 1? --Percy
|
|||||||||||||||||||
Percy Member Posts: 22504 From: New Hampshire Joined: Member Rating: 4.9 |
Hoot Mon writes: Yes, Percey, I went there, but I haven't run the generator yet because I was disappointed by this claim... Why would you want to run the generator? Without any selection mechanism it isn't emulating evolution, so what would be the point of running it? To the evolutionists here the missing selection component reveals the How to Do Your Own Darwinian Evolution Experiments with the Random Mutation Generator webpage to be both wrong and naive. Look at the section nearly halfway down that has the heading There's an Even Bigger Problem. He says:
Let's say you want the word Brown to evolve into the word Black. Shouldn't be too hard, should it? Only four letters need to change after all. But even if you could get the mutations to concentrate just on those four letters, you'd still get a mis-spelled word, which natural selection would eliminate before it ever evolved into the correct spelling. Let's try it, one mutation at a time: Brown > Brorn > Brorb > BrorW > qrorW > qKorW > qKoJW > qKoyW > qFoyW > qjoyW > qjTyW Notice that selection is missing. Without any selection mechanism to cull "mutations" that advance away from the goal instead of toward it, of course the goal is never reached. In essence the webpage's author is claiming to have developed a genetic algorithm program using words as the organisms. But genetic algorithms must have both mutation *and* selection. His program has no selection mechanism and so is not an implementation of a genetic algorithm, and so has no correspondence to actual evolution. A simple selection mechanism for evolving Brown to Black would be to add all the letter differences to provide a single number. Those offspring (there must be multiple offspring in genetic algorithms, another problem with the author's approach) that have the smallest total differences from Black are selected to produce to the next generation. Notice that this is a very weak selection algorithm. I think I can write such a program in about five minutes, so let me give it a try... Okay, here's the program output:
Brown to Black Genetic Algorithm Program with extremely weak selection writes: nclx65% perl ga.plGeneration 1: Brown Generation 2: Bhown Browg Bwown Bxown BrFwn Generation 3: Bhogn Bxoin BwoYn Bkown BroTg Generation 4: Bhfgn Bhodn Boogn Bnoin Bhohn Generation 5: Bhfdn Bmfgn Bhfgs Bhocn Bhfgu Generation 6: Bhfdn Bhfgk Bhfdf Bifgn BhYdn Generation 7: Bhagk Bhddn Bhcdf Bhfdh Bifgl Generation 8: Boagk Bhddk Biagl Bmcdf Bhagi Generation 9: Biddk Biafl Boagl Bhddk Bmcdg Generation 10: Bhbdk Bmcdi Bhddk Biddi Boagm Generation 11: Bhbck Bmcdl Bhbbk Brbdk Biddm Generation 12: Bhack Bhbbk Bibdm Bmcdo Biddm Generation 13: Bkbbk Fhack Brbbk Bidcm Biddj Generation 14: Bkbbk Bkbbm Brbbk Bidcm Ciddj Generation 15: Bkabk Ckbbk Bkbbm Bpbbk Jkbbk Generation 16: Dkbbk Bgabk Bpbbk Bkabf Ckbbh Generation 17: Bpbbk Bpbak Bkadf Ckbbh Dpbbk Generation 18: Dnbbk Bpbbj Bpbbj Bmadf Dpbbk Generation 19: Bmadh Bmadn Dobbk Bpbdj Bpbbl Generation 20: Blbbl Bmadh Dkbbk Bjadn Bmadq Generation 21: Blbbl Bmadm Bmach Ckbbk Bldbl Generation 22: Bmadj Blach Ckbdk Ckbbm Cibbk Generation 23: Blach Blcch Bibbk Bqadj Cpbbk Generation 24: Blacm Blach Blacf Blcch Bgbbk Generation 25: Bkach Blcci Blace Bpach Blaim Generation 26: Bkcci Bkcci Bkaeh Boach Blccg Generation 27: Boacl Dkaeh Bfcci Bdach Bccci Generation 28: Boacm Bjach Boacg Dkaei Dkaeh Generation 29: Bmacg Bjacg Bjcch Dkaeh Dkaeh Generation 30: Bmacg Bmacf Dmacg Bmacq Bmacr Generation 31: Dmacg Dmabg Dmccg Bmaar Bmaar Generation 32: Bmabg Dmabm Dmacg Diacg Fmacg Generation 33: Bmabg Cmacg Fmabm Dnacg Bmdbg Generation 34: Bmacg Cmabg Cmacp Bmdbn Bhabg Generation 35: Bmabg Bhabm Cmaeg Bhabh Ciabg Generation 36: Bhabm Ckaeg Bhabh Cmbeg Ciabg Generation 37: Bhabk Ciabk Bhabm Bhabo Bhaah Generation 38: Cjabk Biabm Bhabj Bjaah Beabk Generation 39: Ckabk Cjabk Biabk Bqabk Ahabj Generation 40: Clabk Boabk Aiabk Biaek Fkabk Generation 41: Alabk Bkabk Clbbk Boabk Aiadk Generation 42: Aladk Amabk Alaak Boabk Boadk Generation 43: Alaal Alafk Biadk Anaak Amabn Generation 44: Alaak Alafk Alaah Anaaj Amabn Generation 45: Amaaj Anabj Aiaak Aldak Agaak Generation 46: Anaej Aiaak Aldek Aoaak Aldal Generation 47: Alaak Aldal Algek Aieak AlZek Generation 48: Aleak Cldal Algbk Algek AlZek Generation 49: Claal Alcbk Albek Alcek Cldak Generation 50: Alcck Claal Albfk Dlcek Alcbh Generation 51: Alcck Dlaek Dlcck Albfh Alcce Generation 52: Aldck Dlahk Albfm Dlccg Dlchk Generation 53: Albcm Ajdck Clbfm Dlccn AlZck Generation 54: Aldcm Ajdck Clafm Blbfm Djdck Generation 55: Agdck Ajdfk Claff Djfck Dpdck Generation 56: Amdck Ajdbk Ajffk Cnaff Drdck Generation 57: Ajdck Ajcbk Gmdck Amhck Imdck Generation 58: Ejdck Ajebk Gmdck Cmhck Gndck Generation 59: Gldck Ajebi Gndck AjYbk Cmhch Generation 60: Cmbch Endck Endck Gldbk Fndck Generation 61: Andck Cibch Fmdck Gmbch Fndcl Generation 62: Ajdck Andck Bndcl Cibcm Andak Generation 63: Andck Aodck Andcl Anddk Bodcl Generation 64: Boacl Bkdcl Andck Apdck Bodbl Generation 65: Boack Boacl Bkdcl Aidck Cpdck Generation 66: Coack Bkdck Bkdcl Boacg Bfacl Generation 67: Aoack Bkdbk Bndck Bkdcl Bndcl Generation 68: Bkdak Dkdcl Bkdch Aeack Bhdbk Generation 69: Bkdbk Dkddl Ekdcl Ceack Bkdan Generation 70: Ekacl Dkdcl Bkdbh Ekdcl Bkdbo Generation 71: Ekacl Dkdcl Epacl Eeacl Bkdgh Generation 72: Akacl Emacl Ekagl Ekecl Egacl Generation 73: Amacl Akacp Epacl Ekafl Egacl Generation 74: Akacm Bgacl Amacd Akaap Enafl Generation 75: Akabm Akacf Bracl Enael Akfcm Generation 76: Bracl Anacf Braci Enael Enaem Generation 77: Bnaci Bracl Dnael Enaek Brbcl Generation 78: Bmaci Emaek Fnaek Dhael Dncel Generation 79: Bnaek Fnadk Doael Emdek Hmaek Generation 80: Bnaek Biaek Bmdek Fnadk Fnadk Generation 81: Biaak Bhaek Bmdel Bnaik Eiaek Generation 82: Biaak Biaek Bhaak Bmdbl Bmdel Generation 83: Bjaak Bmdcl Biaai Btaak Bmdil Generation 84: Bmccl Bmdcj Bjagk Bpaak Ejaak Generation 85: Bjack Blccl Elaak Bjagk Bpaak Success! Here's the final generation: Generation 86: Black Blbcl Bjack Bjaak Bqack If I improve the selection algorithm to be stronger, such as a simple count of the number of correct letters, then it takes fewer generations:
Brown to Black Genetic Algorithm Program with extremely strong selection writes: nclx65% perl ga.plGeneration 1: Brown Generation 2: Bxown BroKn Brywn Bgown Brhwn Generation 3: Bxown BxoBn BVown BxowJ BxRwn Generation 4: BTown BxowY Bxoun BxoWn Bxosn Generation 5: BToNn BToww BTowD BTojn BTokn Generation 6: BToIn BTozn BUoNn BRoNn BNoNn Generation 7: BTogn BToIK BTojn BOoIn BaoIn Generation 8: BTogk Byogn BTogb BMogn BTogm Generation 9: BVogk BTXgk BTolk Blogn BMocn Generation 10: Blogk BPogk BVosk Bdogk BVokk Generation 11: BloSk Blolk BlNgk Blojk BloWk Generation 12: Block BlGSk BloJk Blouk BloIk Generation 13: BlVck plock BloMk Xlock Bloek Success! Here's the final generation: Generation 14: Black Block BlLck BlEck BlVWk Here's the program, it's in perl:
Program Code writes: #! /dv/tools/bin/perl use strict; my $originalWord = "Brown"; my $targetWord = "Black"; my $numOffspring = 10; my $numSelectedOffspring = 5; my $maxGenerations = 100; if (length($originalWord) != length($targetWord)) { print ("Length of orginal word ($originalWord) and target\n" . "word ($targetWord) are not the same, this program\n" . "can not yet handle different word lengths.\n"); exit; } my $wordLength = length($originalWord); my $letters = "abcdefghijklmnopqrstuvwxyz" . "ABCDEFGHIJKLMNOPQRSTUVWXYZ"; my @Letters = split(//, $letters); my $numLetters = @Letters; my @ThisGeneration = ($originalWord); PrintGeneration(1, @ThisGeneration); for (my $gen=2; $gen<=$maxGenerations; $gen++) { my (@NextGeneration, @SelectedNextGeneration); foreach my $word(@ThisGeneration) { for (my $i=0; $i<$numOffspring; $i++) { my $offspring = $word; substr($offspring, int(rand($wordLength)), 1) = $Letters[int(rand($numLetters))]; push(@NextGeneration, [Evaluate2($offspring), $offspring]); } } @NextGeneration = sort {($a)->[0] <=> ($b)->[0]} @NextGeneration; for (my $i=0; $i<$numSelectedOffspring; $i++) { push(@SelectedNextGeneration, $NextGeneration[$i]->[1]); } @ThisGeneration = @SelectedNextGeneration; if (Success(@ThisGeneration)) { print "\nSuccess! Here's the final generation:\n\n"; PrintGeneration($gen, @ThisGeneration); last; } else { PrintGeneration($gen, @ThisGeneration); } } exit(0); ################################################## sub PrintGeneration { my ($gen, @Generation) = @_; print "Generation $gen: " . (join " ", @Generation) . "\n"; } # Uses count of number of correct letters, higher is better sub Evaluate1 { my ($word) = @_; my $sum = 0; for (my $i=0; $i < $wordLength; $i++) { if (substr($word, $i, 1) eq substr($targetWord, $i, 1)) { $sum++; } } return $sum; } # Uses summation of differences of letter ordinals, # lower is better sub Evaluate2 { my ($word) = @_; my $sum = 0; for (my $i=0; $i < $wordLength; $i++) { my $diff = ord(substr($word, $i, 1)) - ord(substr($targetWord, $i, 1)); if ($diff < 0) { $diff = -$diff; } $sum += $diff; } return $sum; } sub Success { foreach my $word(@_) { if ($word eq $targetWord) { return 1; } } return 0; } --Percy
|
|||||||||||||||||||
Percy Member Posts: 22504 From: New Hampshire Joined: Member Rating: 4.9 |
Wounded King writes: I'm pretty sure Perry would categorise all of those intermediate changes as 'fatal' mutations, even though you have done the first thing he said was impossible. I agree, but I thought it important to go beyond just claiming "Selection is missing" by showing how selection can be so easily included. Adding a requirement to the selection mechanism that all surviving offspring be real words is extremely simple conceptually, assuming a spell-check routine. But requiring that all offspring be legal words means there could be many target words to which no evolutionary path is possible if word lengths can't change, so my program would have to be modified slightly to accommodate the possibility of addition/subtraction of letters. I hope we don't spend any more time on that webpage, even though it's the topic of this thread. It's so bad in places it isn't even wrong. Untangling his confused thinking for an audience unfamiliar with the topic would be a considerable challenge. --Percy
|
|||||||||||||||||||
Percy Member Posts: 22504 From: New Hampshire Joined: Member Rating: 4.9 |
DigDug Master writes: If anyone had bothered to run the generator you would see there is a natural selection element. You're right! There's a "Select" button which saves away the current state of the text so that you can easily back up to an earlier stage by clicking on the adjacent "Revert To Selected Text" button. This is "manual selection", not "natural selection", and it's there so that you can repeatedly mutate-revert-mutate-revert-etc until you get a mutation you like. This manual approach is unlikely to make much progress because you're playing with a tiny population of one organism (one sentence), and the organism can have only a single offspring. You really need a computer programming carrying out the steps automatically so that you can have multiple organisms each of which can have multiple offspring (think of a school of fish, for example, where there can be hundreds of fish, and each year each male/female pair produce hundreds of fertilized eggs) to get any meaningful results, see my Message 74.
Ironicly, even though it seems nobody agrees with the site, you would all rather bicker amongst each other then seriusly attempt to reason and explain why it's inaccurate. I'm an open minded person, why not dignify the subject with some clear answers. Sorry, I thought we were being clear. Do you have any specific questions? --Percy
|
|
|
Do Nothing Button
Copyright 2001-2023 by EvC Forum, All Rights Reserved
Version 4.2
Innovative software from Qwixotic © 2024