Monday, March 20, 2017

Correcting the correction of a video about evolution

Charlie McDonnell is the author of a book called Fun Science: A Guide To Life, The Universe And Why Science Is So Awesome. He made a video on misconceptions about the theory of evolution (see below). Sally Le Page (below left) is an evolutionary biologist working on her Ph.D. at Oxford (UK). She noticed a few problems with the McDonnell video so she made one of her own to correct the misconception in the first video. Now it's my turn to correct the misconception in the video that corrects the first video!

Sally Le Page highlights six misconceptions in the McDonnell video. She points out that none of them are very important—they are "little niggles"—but she still thinks a comment is necessary. (I agree.)
  1. The animal in the video is not a monkey, it's an ape!
  2. Evolution is not the same as progress. Trees should not have humans on top. (I could add an additional nitpick—don't use a nineteenth century tree to illustrate evolution.)
  3. Modern bananas are actually designed! They have been specifically bred by artificial selection to produce a tasty treat.
  4. Fish don't have better eyes than mammals.
  5. Evolution is NOT just about survival. It's survival AND reproduction. Modern evolution includes inclusive fitness (her thesis topic). Inclusive fitness is when you help relatives to reproduce.
  6. Sequence similarity doesn't mean anything, those statistics are meaningless. "Are we including the junk DNA. The DNA that we don't know what it does but we think has a really important purpose but we haven't found out that purpose yet so we just call it 'junk'."




You can probably guess that it's misconceptions #5 and #6 that I want to correct.

Sally Le Page is studying at Oxford. She's a fan of Bill Hamilton and Richard Dawkins. It's safe to assume she's a confirmed adaptationist. The problem with the first video is that it focused on the "Theory of Evolution" and on natural selection. I would have mentioned evolution by accident and random genetic drift just to make sure my audience didn't equate evolution with natural selection.

If you are going to correct the original video then it seems very strange to pick on inclusive fitness as the key concept that was omitted. By doing that, you are emphasizing natural selection over drift and that just adds to the original misconception. Inclusive fitness is a insignificant factor in evolution in 90% of all species because individuals don't even know who their relative are. (Think bacteria and protozoa.) In the rest, mostly animals, inclusive fitness only accounts for fixation of a small percentage of the alleles that are fixed by natural selection. Furthermore, alleles fixed by selection are only a fraction of those fixed by random genetic drift.

I'm afraid that Sally Le Page has missed the real problem in the original video and only made the misconception worse by bringing up inclusive fitness.

The last point (#6) is complicated. Charlie_McDonnell said in the video that we share 50% of our DNA with bananas. This is clearly incorrect. I don't know what the exact number might be but let's do a back-of-the-envelope calculation. Let's say that bananas and humans have 10,000 orthologous protein-coding genes and let's say the coding regions are 50% identical in amino acid sequences. That would account for less than 0.5% of the genomes. Adding in a bit of other genes and functional sequences won't get us over 1%. The rest is mostly junk DNA and it's unlikely there is any significant sequence similarity. We certainly don't share 50% of our DNA with bananas.

However, Sally Le Page goes too far when she dismisses all such comparisons as meaningless. When we say the genomes of human and chimps are 98% identical in sequence and humans and macaques are 94% identical then that actually means something [How do you explain the differences between chimpanzees. humans, and macaques?]. We can also say that orthologous human and banana genes are 50% similar whereas human and bacterial genes are only 30% similar. That means something.

Like all good adaptationists, Sally Le Page is probably skeptical of junk DNA because the presence of so much unselected DNA goes against her worldview. That's why she says that junk DNA isn't really junk—it's just DNA that has a purpose we haven't yet discovered. (This is the "dark matter" point of view.)

She is entitled to her opinion but she is perpetuating a misconception if she fails to mention the views outside of Oxford. If you are really interested in educating the general public about evolution then you'd better make sure you mention controversial ideas and make sure you aren't promoting your personal opinion as fact. In this case, the correct statement would be that less than 10% of our genome is conserved in other species and the rest may or may not have a function. Many scientists believe it is junk and many believe that it has a mysterious unknown function.


58 comments :

  1. You have your obsessions, I have mine, so I will deal with #1: apes are monkeys. Humans are apes. Humans are monkeys. Humans, other apes, and other monkeys are primates. I could go on for a long time. Hey, look! It's groups within groups! They both manage to make that mistake, though in different ways.

    She also seems to think that circle trees are the "modern" sort of tree, when they're just one graphic approach to drawing trees with identical content.

    Why would anyone say that sequence similarity is meaningless? I see that what she might really mean, but never manages to say, is that there are different measures of similarity. I'd agree that some of them are close to meaningless. Others are meaningful.

    ReplyDelete
  2. Groups within groups are all good and well, but if you say humans and other apes are monkeys, you are basically saying that monkeys are just another name for haplorhines. But that isn't how the term "monkey" is normally used -- instead it is a group which expressly excludes the apes because they don't have tails. Yes, that doesn't really make any meaningful biological sense, but that's exactly why "monkey" is only a layman's term that should be avoided.

    ReplyDelete
    Replies
    1. But you could make the same objection to "ape", another layman's term. Both "ape" and "monkey" are ambiguously defined in popular understanding, and an evolutionary definition is both possible and pedagogically useful. I choose to exercise that option in both cases, and it's certainly more sensible, at least in a discussion of evolution, to say that humans are monkeys than it is to say they are not. Paraphyly, at least in a discussion of evolution, is a bad thing.

      Delete
    2. Both ape and monkey are paraphyletic groups which are not "correct" (i.e. monophyletic) in a scientific arena. This is one of the trade-offs of making science more accessible to the layperson, sometimes you have to use colloquial terms that aren't entirely accurate.

      Delete
    3. I think it's highly useful to tell people that humans are apes and apes are monkeys. Nothing says "evolution" better.

      Delete
    4. Paraphyly, at least in a discussion of evolution, is a bad thing.

      And illegal in 27 states.

      Delete
    5. John,

      I agree with much of what you say. We also have to recognize where those labels can get confusing. For example, we may define a hominid transitional fossil as having a mixture of human and ape features, but if humans are already apes this type of definition doesn't make much sense.

      Delete
  3. Inclusive fitness is a insignificant factor in evolution in 90% of all species because individuals don't even know who their relative are. (Think bacteria and protozoa.)

    I don't believe this is required. Inclusive fitness also applies to microbes whenever there is some spatial pattern that keeps related cells together.

    ReplyDelete
    Replies
    1. Yes, what Corneel said is correct. Larry shouldn't criticize a concept he doesn't fully understand. Inclusive fitness pervades all aspects of life--sets of related cell lineages, protozoa, individuals, vegetative propagation, coral reefs--all of these entities can evolve due to differences in relatedness patterns. It is not a trivial issue and does not rely on who has a conscious understanding of who their relatives are.

      Delete
    2. I was wondering whether quorum sensing systems in bacteria provide a mechanism for inclusive fitness.

      Delete
  4. When we say the genomes of human and chimps are 98% identical in sequence

    Does this 98% include the junk DNA? If yes, then what is the similarity if we ignore junk?

    ReplyDelete
    Replies
    1. Actually, the similarity is 98.7%, and it includes the junk DNA. If you ignore junk the similarity is greater. I don't have a figure for that exact measure, but if you count only protein-coding exons, which are around a fifth of the non-junk DNA, the similarity is 99.5%.

      Delete
    2. Actually, the similarity is 98.7%, and it includes the junk DNA. If you ignore junk the similarity is greater. I don't have a figure for that exact measure, but if you count only protein-coding exons, which are around a fifth of the non-junk DNA, the similarity is 99.5%.

      It must be an evolutionary miracle that chimps and humans don't look the same with such precise accuracy of their genetic similarities...

      Unfortunately they (the rules of interpretation) have been decided already by the "science bullies"...If you are in control, you must be right... and even if you are wrong, "you must be right by privilege..." at the time... and time can change very quickly...

      BTW: My father and I are 93.7 % similar genetically... Is he a primate? Harshman, what's your assessment now?

      Delete
    3. By what measure are you and your father 93.7% similar? I will try to address your ignorance if you explain it further.

      Delete
    4. I have to assume that Jass and his father have had DNA tests done by a company like 23andme. If so, here is how they derive their calculations:

      "The One-to-Many tab allows you to compare a single person with anyone with whom you are sharing genomes. You can make the comparison across all the SNPs on our chip, which gives you a rough estimate of genetic similarity for this 0.03% of the human genome, or you can look only at genes connected to a particular trait by selecting that trait from the list on the right."

      So it is based upon 0.03% of the genome... I suspect that Jass does not understand this.

      Delete
    5. Sorry, here is a link to the complete 23andme overview from which this was copied:
      https://customercare.23andme.com/hc/en-us/articles/202907130-Gene-Comparison

      Delete
    6. By what measure are you and your father 93.7% similar? I will try to address your ignorance if you explain it further.

      Does it matter? Are there more than one measure? I thought there was a consensus about this shit?

      Delete
    7. There are many possible measures of genetic similarity with many uses. You know nothing. If you give me more info about where you got this measure, I can try to figure out what yours means.

      Delete
    8. Since you seem to know everything, just tell us why you and the rest of........like you can be trusted?
      Lying seems natural to you....

      Delete
    9. Joeys harasmant movie must be fora good cause...eliminating the competiton...one by...one.

      Delete
    10. Jass, it is contemptible to accuse someone of lying and not present a single shred of evidence in support of this serious accusation.

      Delete
    11. Jass is a cum laude graduate of the Trump post-truth academy.
      As a graduate you never have to provide one shred of evidence.

      Delete
    12. Jass, it is contemptible to accuse someone of lying

      I think the adjective should actually be sad.

      Jass saw a number, has no earthly idea what it really means, but feels he cannot admit that here. So when someone who could actually explain things offers to do so, having no way to evaluate its bona fides, instead of being informed he lashes out in suspicion and anger.

      Going through life ignorant (even though voluntarily so), suspicious and angry cannot be a piece of cake, and thus I think Jass deserves our pity more than our contempt.

      Delete
  5. Well, if that 0.03% of the genome were an unbiased sample, it would be an estimate of total genetic similarity. Of course it isn't. It's actually a sample of known SNPs that are polymorphic at a reasonable frequency, which would of course greatly overestimate genomic divergence. It would certainly be no surprise for a father and son to share alleles at 93.7% of those SNPs. If anything, it would seem to suggest a bit of inbreeding.

    I note that this (except the inbreeding bit) is actually explained well on that page.

    Jass, has that addressed your ignorance sufficiently?

    ReplyDelete
    Replies
    1. Actually, I'm a bit interested in the 93.7% number. Presumably 50% identity comes directly from the father and 43.7% from the mother. That implies to me that the mother would be 87.4% identical to the father at the assayed SNP sites. Can anyone clarify that point? That suggests that the parents were related, how closely I would not guess.

      Delete
    2. Presumably 50% identity comes directly from the father and 43.7% from the mother.

      Similarity doesn't necessarly imply alleles are IBD. My guess is that many SNPs were not variable in the parents, because they were sharing the most common allele.

      Delete
    3. Unrelated individuals, according to the web site you linked, are generally about 70-75% similar.

      Delete
    4. But Jass is presumably related to his father, no?

      Delete
    5. But Jass is presumably related to his father, no?

      Yes, unless the mailman is a very close relative, for example his father's brother, and the community is very isolated.

      Delete
    6. OK, I understand your point now. Partially, that is. I'm still not sure how you calculate the expected similarity between Jass's parents from that between him and his dad.

      Delete
    7. A close-knit religious community (over several generations) might explain it, or recent ancestry from such a community.

      Delete
    8. One other thing that might explain it is a simple mistake on Jass' part, confusing similarity in a particular aspect of genetics with similarity across the entire .03%. See the Circadian Rhythm example at the 23andMe page.

      Delete
    9. lutesuite,

      Simple calculation. Exactly half of Jass's SNPs come directly from his father, and those will be identical by descent. So he's guaranteed that 50% similarity regardless of the mother or the distribution of SNPs in the population. Exactly half (the other half) of his SNPs come directly from his mother. That half accounts for another 43.7% similarity between Jass and his father, and since it comes direct from his mother, must also equal the similarity in that half between father and mother. That is, 50% of Jass's SNPs are 87.4% similar between his mother and his father. Assuming those SNPs are not a biased sample of the mother's full set of SNPs, and I don't see how they could be, then the mother's full SNP set is 87.4% similar to the father's.

      Of course I ignore mutation here, but that should be negligible.

      Delete
    10. Thanks. I'll have to think that thru a bit.

      Delete
    11. Unrelated individuals, according to the web site you linked, are generally about 70-75% similar.

      Ah, I didn't see that part. Then I'm betting on judmarcs suggestion that Jass was just reporting a diiferent number. The similarity strongly depends on the SNP panel being used.

      Delete
    12. All these numbers about what percent related different relatives are describe different things. At the DNA level, I am 99.9% similar to another random human. Most of these differences have low population frequencies (as they should for neutral variations). If I have one of these rare variants, at that site this rare allele will show up close to 50% of the time in my brother.

      Based on that what is the genetic similarity of me to my brother? 50% or 99.9%? Both numbers are useful, in different contexts. If we are asking about the dynamics of a rare allele that predisposes me to some altruistic behavior, it is the 50% figure we want. If we want to know about how similar DNA sequence will be at a random site, the 99.9$ figure is the one we want.

      There is no contradiction between these numbers. They are connected by a formula discovered by Sewall Wright in 1921. If I look at a site in one of my chromosomes that has an allele with frequency p, and consider one of the two copies of that chromosome in a relative that has probability of identity-by-descent of f, the probability that both of us have the allele is

      (1-f) p^2 + f p

      For a sibling, f is 1/4. Not 1/2 but half that as we are only looking at one of the two copies of the chromosome in my relative. For an allele whose frequency in the population is 0.001, the probability that it shows up in a particular copy of the chromosome in my sibling is 0.001 x (1/4) + 0.001^2 x (3/4) or 0.00025075. But for the other allele it is 0.999 x (1/4) + 0.999^2 x (3/4) or .9982507500. Overall it is the sum of these.

      The exact number overall depends on the distribution of gene frequencies of rare alleles. I note that the calculations I am doing here are basically the same as the ones John Harshman was doing, above, just a bit more explicitly,

      Delete
    13. Correcting myself: 99.9$ should obviously be 99.9%. And the probability that I am calculating for me and my sibling is the joint probability, not the conditional probability. I should have typed: "the probability that it shows up both on a particular copy of my pair of chromosomes and in a particular copy of the chromosome in my sibling is".

      Delete
    14. The 1,000 Genomes Project found that the average genome differs by 4.1 to 5 million sites.

      http://www.nature.com/nature/journal/v526/n7571/full/nature15393.html

      I don't know if this in reference to the haploid or diploid genome, but in either case we are looking at much less than a 1% difference between any two human genomes. It is nowhere close to the 93.7% touted before.

      Delete
    15. 4.1 to 5 million sites out of a haploid number of 3 billion sites would give a similarity of 0.99863 to 0.99833. The usual figure previously was 1 difference per 1000 bases between two unrelated human haploid sequences which would give 0.999, the figure I cited above.

      Delete
    16. Maybe folks are over thinking it. Perhaps the 50% figure comes from more of an simple approximation based on large numbers and doesn't deal with coding/non-coding regions, orthology, or genealogical relatedness based on IBD.

      That is, since there are only four nucleotide states: A, C, T, and G, if you randomly generate long long stretches of DNA (with random stringing together of A, C, T, G), they will be identical at 25% of those positions (as the stretches go to infinity, the sequence similarity is 1/4).

      Hence, perhaps they were just being cheeky (I didn't watch the video) in that given the baseline level of sequence similarity between any two creatures is 25%, thus when comparing a human to a banana (which is more related to us than, say, giardia), they just guessed it would be about 50%, since "plants" are "in between" humans and bacteria in terms of evolutionary relationships.

      I'm just speculating as to their reasoning.

      Delete
    17. As Rich notes, another 50% figure is the often-cited 50% similarity between humans and bananas. That was, I think, a figure as to what fraction of loci showed noticeable homology, based on protein sequence. "Noticeable homology" is a far less stringent condition than having 50% similarity in protein sequence, and certainly does not imply 50% similarity in DNA sequence.

      Delete
    18. So in other words, to say that we share 50% of our DNA with bananas is to say that 50% of our protein coding genes have homologues in banana (which may or may not be ~50% similar in sequence). Is that correct?

      Delete
    19. I believe Donald Trump shares 50% of his DNA with a smaller species of banana.

      Delete
    20. @Mikkel: 50% having noticeable homologues can mean that there is no detectable similarity in DNA sequence, but that there is detectable similarity in amino acid sequence. And yes, that is possible.

      Delete
    21. @Joe
      Sure, I understand why, bc of the genetic code with multiple codons for the same amino acid. But I really wanted to know if all that was meant by that saying (we share 50% DNA with bananas) is that about half of all our ~20.000 protein coding genes have homologoues in Banana, so about 10.000.

      Delete
    22. OK, so I decided to actually look up the 50% human-banana figure. Names that popped up in Google searches included Evan Eichler (who has the lab across the hall from me), Dan Hartl (who I have known since our undergraduate years), but neither seems to have actually said that. Nor was it in the Nature paper in 2012 that first published the banana genome. Steve Jones does seem to have said, in lectures, that humans and bananas share 50% of their DNA. Which is quite wrong. However he was in science-popularizer mode at the time, so his statement cannot be considered a primary source.

      The figure gets repeated all over the place. It is sometimes given as sharing 50% of their genes, which could be interpreted as having 50% of proteins showing noticeable similarity of protein sequence. Which might be true.

      But even then, I don't know where the figure comes from. Can anyone enlighten me on the actual source?

      Delete
    23. Joe,

      Let's not fool anybody! You chose to believe something, just because your faith-buddies expect you to and your reputations is at risk if you dared to choose otherwise. You will follow what you need to follow; right or wrong...
      So, your statement means shit....and it does't matter to you does it?

      Delete
    24. Troll's gotta troll. Go get em Jass, you big-old do-gooder of biological wisdom you.

      Delete
  6. You guys are using old fashioned, regular math. You still haven't figured out Intelligent Design Creationist Jesus Math (TM). By this new and improved math, 93.7% of 0.03% = "I ain't no freakin' monkey".

    ReplyDelete
    Replies
    1. Remarkably, the Creationist Jesus Math (TM) has only one axiom. It's from the AIG statement of faith:
      "By definition, no apparent, perceived or claimed evidence in any field, including history and chronology, can be valid if it contradicts the scriptural record. Of primary importance is the fact that evidence is always subject to interpretation by fallible people who do not possess all information." - Answers in Genesis

      And there we have the foundational falsehood of creationism. They assume they're right by definition and all subsequent work is done only in an attempt to support it.

      Delete
  7. British evolution defence are terribly inferior because they have not been dealing with constant creationist attacks. in short selection has not occurred to weed out the poor answers relative to less poor answers.
    Likeable people but history is moving on here folks.

    If eys are poorly done because of a water origin then why not have evolved to a position where a former eye adaptation has vanished? they have everything evolving and so why should the eyes linger behind? AHA its because of a lame attempt to defeat the creationist charge that eyes are too complex to have evolved. Aha. So they welcome a flaw but it defeats the whole point of the glory of evolutionism to turn a fish into a rhino .

    Its still all speculation on speculation as they present it.

    The reason for trees with humans at top is because the victorian elite saw man as the most intelligent and so most evolved. The rest didn't evolve in smarts and so are below us. they got a point and modern evolutionism doesn't.
    I agree worms are as complex as us in body however.
    this from a common design and time could not change such a complex design.

    Yes we have almost like dna with primates as shown by our looks. how else could it be? could we look so alike and not have like dNA? No! Not if there was a creator with a common design for biology!
    Biology is obviously off the same rack. like some guy would do it.
    Evolutionism is demanding a creator must, to justify creation, have made biology in such a random way as to have no rhymne or reason.
    Again lines of reasoning
    At least these bRits show even in europe there is increasing criticism from the public. at last euros are questioning authority.
    What they should demonstrate is the top thre biological scientific evidences for evolution or admit there are not such things.
    That would be a youtube video to change the world.


    ReplyDelete
    Replies
    1. Putting evidence for evolution up on the internet? My God, what a brilliant idea! Only a Robert Byers could come up with such a thought.

      Delete
  8. "It's not a lie if you believe it" -someone once said...

    I often say that ( to some) "...it's not a lie if you want to believe it..." However, it is an unacceptable lie if you want to deceive others because you chose to believe a lie and manipulate others to believe it...

    Since there is no system to make public deceivers accountable to society, anybody can publish any shit he wants under the protective umbrella of "science" and he gets away with that whether it is true or false...with one exception: ID. If they publish something even with experimental evidence that can't be denied, it's always wrong, a lie or deceiving...

    This is what's wrong and I'll fight it to death....

    ReplyDelete
    Replies
    1. Sorry Jass, but you're fighting the wrong fight. Like Don Quichotte you think you're attacking giants but in fact you're attacking something which isn't even there.

      I have to agree with Judmarc:
      Going through life ignorant (even though voluntarily so), suspicious and angry cannot be a piece of cake, and thus I think Jass deserves our pity more than our contempt.

      Delete
  9. Dr. Moran, what are your thoughts on this article that I came across about a possible purpose of junk DNA.

    https://www.labroots.com/trending/cell-and-molecular-biology/5601/transcription-mechanism-sheds-light-junk-dna

    I realize the discovery is in unicellular organisms only. Is this anything new or is it making a larger claim than is appropriate? Thanks.

    ReplyDelete
    Replies
    1. This is a weird situation. In this Paramecium, there is a small germ-line nucleus and a big, polyploid nucleus that runs the cell. The germ line nucleus contains lots of sequences derived from transposons. Probably junk DNA. In the macronucleus, those useless sequences are removed. But how are they removed? By using them as templates for making little RNA’s that will be involved in cutting them out! So are those sequences useless junk? Or are they useful because they’re involved in the process of removing the useless junk – which is themselves?

      I love how real things have no respect for our need for make simple categories.

      Delete
    2. Isn't it fair to say, bwilson295, that the sequences in the germ line nucleus are not junk but the same sequences in the polyploid nucleus are junk as they are not necessary? In fact at some point shouldn't the editing function cease to work? After all, in the germ line nucleus the sequences will be under selective pressure as they are useful, so they are unlikely to change rapidly, but in the polyploid nucleus the sequences would not be under any conserving selection. So eventually the sequence in the germ line nucleus would not match the sequence in the polyploid nucleus sufficiently well to result in the excision of the sequence. Or so I would think, at this point, before I hear your reply.

      Delete
    3. The sequence in the germ line nucleus will match that in the bigger polyploid nucleus because the bigger nucleus is made from the germ line nucleus each generation. Sorry I didn't make that clear.

      It's possible that you are right. The article, which focused on a small part of this process, didn't address potential usefulness of the sequences except for removing themselves. In general, transposons exist for their own good -- to reproduce themselves like parasites in the genome of something else. And transposons mutate, so most copies become useless junk. However, there are some cases where they host puts a transposon to work, so to speak. Because of what's known about transposons in general, I would expect these sequences to be useless or mildly harmful to the Paramecia, but that might not be true.

      Delete