Creeping Mendelism!

I dig into one of my YouTube comments, and also deplore a failing of genetics teaching. We have got to get away from Mendel! It’s nothing personal, his ideas about genetics were primitive (but useful!) and are now outmoded.

Transcript below the fold.

I’m in slo-mo mode today. I’m on drugs to deal with back pain! I think I still have enough brain to deal with Jordan Peterson fans, though. So I found this comment on last week’s video and thought it would be pretty easy to handle, even with my half my brain tied behind my back, or whatever inept phrase Kent Hovind uses.

The issue is that I ripped into Jordan Peterson for his foolish idea that if he takes the right kinds of drugs, he can turn his brain into a microscope and see molecules of DNA. He presented no evidence that he can do this. I guess he needs to learn that facts don’t care about your feelings.

I got a surge of stupid traffic from this: people logging into tell me that drugs are wonderful, that I didn’t understand the meta concept Peterson was discussing, calling me stupid, that I was missing the context behind his claim, etc. They were trying so hard to rationalize the patent idiocy of Peterson’s claims. I’m so sorry for them!

And of course, one of the more common arguments was an ad hominem — and surprise, I know that that means, unlike most of the people throwing the Latin around. It does not mean insulting someone. It means using a different, unrelated component of my character to dismiss my ideas on the subject at hand. Here’s an example of someone announcing that because I’m a so-called Social Justice Warrior, they can just ignore what I said about Peterson’s magic DNA eyes. They say,

I am not going to talk much about Peterson, but here is my problem. PZ Myers is supposed to be a scientist and yet he let’s social justice which has nothing to do with Science leak in.

There’s a remarkable assertion: science, or Capital-S science, can’t have anything to do with social justice. Actually, it can, if conditions have empirical, measurable consequences, so that’s kind of nonsensical. Even if it didn’t, scientists are human beings, and we can care about whether other people are being treated fairly. This isn’t the objection he thinks it is. But he rushes to tell me he’s a good person anyway!

I help a Transgender person overseas and help feed him and fix his bike, so this isn’t about hate or anything, but pronouns and having many sexes is against the Scientific data.

Oh. The “I have a black friend” argument.

It’s very nice that they have helped a transgender person, but now I’m confused. You used the pronoun “him”, so are they a transgender man, or are you misgendering a transgender woman? I only ask because you say pronouns are against the scientific data, and there you go, using pronouns like a non-scientist. Or a normal human being.

They continue digging their hole.

It’s more like a problem with the mind itself and social justice should not be mixed with Science.

So Capital-S scientists shouldn’t study the mind or society? Peterson is a psychologist, maybe you should go have a conversation with him about how he can’ claim to be a scientist (I’d agree, by the way, but psychologists can be legitimate scientists, you know). There’s a whole category called the Social Sciences, which includes psychology and sociology. Who are you to define what science is or isn’t?

Let’s wrap this up.

This is why I am upset, because if you are a Scientist, you should have NOTHING TO DO WITH STUFF THAT’S NOT SCIENCE AT ALL, it’s more pseudoscience than actual real science.

Disappointed in you PZ, I thought you would be better than that.

This is really shameful and I think that’s worse than whatever Peterson is going on about.

OK, here’s what I find annoying myself, and it’s something I see a lot of in the comments, especially when I say anything about transgender issues. The dumbest, most ignorant people show up to tell me that I don’t know the science, THEY know the science far better than I do. All they really know is a few bits and pieces of some poorly understood genetics that they recite to bolster their prior biases. They also ignore the injection of politics or sociology that they agree with into the pseudo-science that his preferred characters, like Peterson, will babble about.

It’s ridiculous that he finds a scientist advocating for social justice to be more objectionable than a mad, deluded kook promoting hateful bigotry.

But I’m trying to figure out why he thinks “Scientific data” is against pronouns or a more complex perspective on sex. And I think I know how they come to this conclusion, and it’s a common problem in genetics education. Too often, early genetics instruction in the public schools begins with Mendel, and ends with Mendel, leaving students thinking that genetics is relatively simple and deterministic. I suspect that many of the listeners of this video think they know the basics of genetics: there are dominant and recessive genes, they are sorted and recombined in meiosis and fertilization, and which one you express is a binary decision. The canonical story of the Y chromosome fits perfectly with this simplistic idea: Y is the male chromosome, if you inherit one, you are male, and if you don’t get one, you’re female, and that’s it.

That’s the problem. Mendel’s work was a preliminary insight into a possible mechanism; his results from pea plants were useful and interesting and once they were rediscovered, promised a way to analyze inheritance with mathematics, probability theory, and statistics. If you stop at Mendel, though, you miss all the interesting complications that emerge later, and you acquire a primitive and rigid notion of how inheritance works.

I’ve been calling it creeping Mendelism, the way a few experiments with pea plants have crept into the public consciousness and taken over. I’m not the only one who has noticed! Here’s a paper that was published a few years ago that says the same thing:

Twenty-first-century biology rejects genetic determinism, yet an exaggerated view of the power of genes in the making of bodies and minds remains a problem. What accounts for such tenacity? This article reports an exploratory study suggesting that the common reliance on Mendelian examples and concepts at the start of teaching in basic genetics is an eliminable source of support for determinism. Undergraduate students who attended a standard ‘Mendelian approach’ university course in introductory genetics on average showed no change in their determinist views about genes. By contrast, students who attended an alternative course which, inspired by the work of a critic of early Mendelism, W. F. R. Weldon (1860– 1906), replaced an emphasis on Mendel’s peas with an emphasis on developmental contexts and their role in bringing about phenotypic variability, were less determinist about genes by the end of teaching. Improvements in both the new Weldonian curriculum and the study design are in view for the future.

This is not the same as saying that Mendel was totally wrong! Mendel defined a simple starting point, but the progress of genetics since has since eclipsed his preliminary ideas. I suspect that most of you are wondering what in your education is obsolete. So let me go over in the most basic way what Mendelism says.

Mendel worked backwards to infer the exist of what he called unit factors in gametes. So an organism has a set of traits which I’ve illustrated with some colored solids; in Mendel’s case, he identified a set of traits in pea plants, such as the color of the seeds or flowers, or the height of the plant, or other properties of the plant. Then he did a set of carefully defined crosses, followed by meticulous accounting of the progeny, to infer correctly that each plant had pairs of alleles for each trait, and that some alleles were dominant (that is, always expressed) or recessive (that is, not expressed if a dominant allele is present), and that they were distributed to progeny randomly.

Does this sound familiar? It should. You might have gotten this in basic biology classes when you were about 12. The one thing you might have learned beyond what Mendel knew is that these genes were encoded in a molecule called DNA.

The unfortunate implication that you may also have absorbed is that there is a one-to-one correspondence between genes and traits. That’s the problem. When your education is arrested at that point, you start thinking in terms of “genes for X”, where X is cancer, or obesity, or eye color, or sexual preference. That means you don’t know about all the cool stuff that comes later.

There isn’t a one-to-one correspondence. Each gene contributes to multiple phenotypic characters, a phenomenon called pleiotropy, and each phenotypic property is established by multiple genes, they are multigenic. People are now talking about the idea of omnigenicity, or that every gene makes some contribution to every trait. This means that the idea of a gene FOR a specific purpose mostly doesn’t work.

Another example: you can’t say that a Y chromosome makes you male. The Y chromosome has one strong effect in switching genes off and on in autosomes, and even the X chromosome, that are essential for male development.

The molecular biology revolution from the 1950s also tells us that the DNA is transcribed into RNA, which is exported into the cytoplasm. Further, most, but not all, of the RNA is translated into proteins. This is the only way genes can act, by sending out molecular proxies in the form of RNA and protein that do the actual work of the cell.

To wreck those nice linear genes-to-phenotype arrows further, the cytoplasm is a stewpot of interactions and activity. Proteins are communicating with and modifying the activity of other proteins, and are feeding back to regulate the expression of genes. They also modulate themselves with feedback and feedforward mechanisms. You simply can’t draw a straight line from a gene to its effect on the phenotype! There are always side effects. It may also be that the mutant feature that leaps to the eye of the geneticist is one of the minor effects of that gene, and there is a more important primary effect that we just don’t notice.

Finally, there’s the immense effect of the environment, a parameter Mendel didn’t even consider. The expression of genes is affected by environmental factors, and also, our biology feeds back and modifies the environment.

What this means is that genes do not determine fate in any simple way; every phenomenon, such as cardiovascular disease in this diagram, are the product of a causal web, of which genetics is an important, but not the only, part. Mendel had intentionally executed a thorough job of experimental reductionism, eliminating all the variables that might have affected his results other than the ones he chose to test, which is good practice in a controlled experiment, but that also artificially exaggerates the effects of that one variable.

And that’s how I end up with a bunch of commenters who can insist that they know the science better than I do — the first, simplest knowledge they have about genetics has instilled in them an absolute moral certainty that genes are deterministic, and that anyone who talks about complexity and subtlety and the importance of other factors must be a bad scientist. They’ve got everything backwards!

Anyway, now you know where I’m coming from. Maybe you also understand a little better why the people who claim genetic differences drive differences between the sexes, between races, and between nations. They’ve got a cartoon version of genetics in their heads.


  1. DonDueed says

    Popular science reporting amplifies this problem with every headline that hypes the discovery of a “gene for” this or that.

  2. birgerjohansson says

    And let us not forget that many genes have more than one effect, not all of them positive for the organism.
    And genes get methylated, which makes things even more weird.
    And the diversity of commensal organisms in your gut apparently is very important for development during the first years (malnutrition suppresses the diversity), so molecules that leak through membranes into the body have co-evolved with human DNA to make development work the way it should.

    It is surprising development is robust enough to produce functioning adults!

  3. StevoR says

    This is why I am upset, because if you are a Scientist, you should have NOTHING TO DO WITH STUFF THAT’S NOT SCIENCE AT ALL,

    How utterly ridiculuous. For starters does that mean someone can’t also enjoy art or discuss politics aor other areas at all? Because, no, just no. (Also what an unscientific unsupported subjective claim to hold!)

    But then also what about Jordan Peterson being a supposed scientist and then supporting religion and, in essence here, mysticism / gnosticism with his drug trip ideas which aren’t scientific as well as his views on the same topics PZ has eg,. Social Justice issues but just in the opposite direction? Reckon he wrote to Peterson saying he can’t rant regressively about the same topics for the same reason? Me neither.

  4. Rob Grigjanis says

    StevoR @3:

    This is why I am upset, because if you are a Scientist, you should have NOTHING TO DO WITH STUFF THAT’S NOT SCIENCE AT ALL,

    How utterly ridiculuous.

    Well, yeah, but it sounds a lot like some atheists talking about scientists with religious beliefs.

  5. wzrd1 says

    @DonDueed, indeed! How often we hear of some blurb about BRCA1 or BRCA2 being “the breast cancer gene”, which is utter bullshit. The gene is part of one system that repairs double strand DNA breaks. One, out of quite a few known and likely, a few more to yet be discovered. DNA is somewhat delicate to begin with, with some gene sequences being downright fragile and incessantly needing repair.
    If that wasn’t the case, we’d likely be extremophiles or well, anything but human.
    Initially, it was thought that there’d be a gene for each function and protein – right until sequencing uncovered that to be not even wrong. A sequence gets mapped in, a distant, possibly entirely different chromosome’s gene might get mapped in next, to make a great honking big protein that incorporates dozens or more protein fragments that became that gigantic monster of a protein. Start going into enzymes and well, therein lies the pathway to madness… ;)
    Of course, most of those commenters fail to even comprehend what an enzyme is, they’ll think laundry detergent and never realize that they’ll not break down food or synthesize peptide hormones without enzymes helping catalyze the reactions.
    I take methimazole, it inhibits the enzyme thyroperoxidase from catalyzing the reactions necessary to bind iodine into thyroid hormones T3 and T4. That, because my immune system is sending trainloads of antibodies to attack my thyroid stimulating hormone receptor. One theory being that the receptors have a mutation that fails recognition by the body, other theories are that it’s a mistaken sensing of an invading organism’s proteins that “look” very similar to the immune system. Frankly, it’s not very relevant, as neither can be corrected easily, if at all, so blocking the process prevents ongoing disease of excessive T3 and T4 production, with most of the T4 being eventually converted into T3.
    They’ll all be looking for the Grave’s disease gene, fruitlessly, as there isn’t one and could never be one.
    Just as cancer isn’t one disease, but a multitude of dysregulating problems that unify in the form of unregulated cell proliferation and an inability for those cells to die when they malfunction.
    They’ll demand finding “The Gene” that causes cancer, scientists know it’s dozens or more that can lead to cancer and hence, there are dozens to hundreds of different types, so seeking one magical cause is a fool’s errand.
    Obviously, since they’ll not like that answer and refuse to learn about what they blather about, they’ll blame the scientist.
    And never blame their priests of ignorance, whose answers aren’t not right, but not even wrong.

  6. birgerjohansson says

    DNA is not, and has never been a “blueprint”. At best, it is a recipe. And each iteration gets some letters altered at random. And the ovens are of very different performance. And some of the ingrediens are not available in the expected amounts.

    And yet, most of us end up with two arms, two legs and a (mostly) functioning head! The various differences -some are left-handed, or gay, or have the heart on the opposite side of the chest- are trivial.

  7. birgerjohansson says

    Fun fact; it is 02:05 in the morning, and the sun is about to rise.
    When PZ has recovered I hope he one day gets to travel to northern Scandinavia in summer (and I hope the rest of you also get the opportunity). Nor are any MAGA hats in sight.
    OK, back to the topic.

  8. Russell P says

    This is one of the best lessons you’ve given. In fact, this has inspired me to re-write how I teach unit 3 in my genetics class. Unit 1 is The Stuff of Genetics (DNA, genes, chromosomes, genomes). Unit 2 is How Genes Work (transcription, translation, gene regulation). Unit 3 is Inheritance (mitosis, meiosis, and the Mendelian stuff). Unit 4 is Genetic Change (mutations, evolution, and cancer). As I’m envisioning it, in Unit 3 I will start with complex traits rather then end with complex traits. We’ll proceed to dissect a complex trait (I need to pick which one) into it’s individual mechanistic components (discussing the impact of each one of several genes plus the environment). Then we’ll work out the basic Mendelian proportions for each of the genes separately (standard Mendelism). Then we’ll put the component genetic effects back together using the sum rule of probability to create the continuous distribution of the character variation. Or something like that–I’ve got the rest of summer to work out the details.

  9. Pierce R. Butler says

    Then he did a set of carefully defined crosses, followed by meticulous accounting of the progeny…

    Did he? Can’t recall where, but someone alleged quite some decades ago that no one trying to replicate the abbot’s pea patch (among countless high school and college gardens) has ever come out with numbers so neat and tidy, suggesting he cheated.

    Or should I file that with the ivory-tower legends?

  10. says

    Biology is not a nice neat computer program. It’s a messy, sloppy hash of what just works. Or what mostly works. As long as 99% of your cells are transcribing properly from DNA -> RNA -> Protein, things generally work. Throw a bit of radiation or other mutagen into the mix and things get really messy really fast. We do NOT live in a clockwork universe designed by a perfect creator. Our cells are doing the best they can to stay alive.

    Free Thought Blogs was being a little weird, I meant to post this comment here.

  11. dragon hunter says

    I think the “Selfish gene” is also a great contributor to this. No the necessarily the book itself, but the oversimplified way in which its main message is transmitted. Even the author repeatedly says that the central idea of that book, is that bodies are just survival machines for genes that are really good at making such machines, and that is why the information on those genes remains virtual unchanged for millions of years. Yet, what modern biology all seems to suggest is the opposite, is that gene transmission mostly results from the collective forces acting on phenotype, which is often not related to individual fitness at all. In such circumstances, even from a purely engineering perspective, having multiple genes involved in a specific trait makes much more sense, as it creates redundancy, which helps deal with random events taking important genes out over evolutionary time.

  12. dragon hunter says

    #10. Yes, this is well known that the numbers are slightly fudged to make them fit the expected proportions better. Although I always thought whether Mendel really could have counted all those peas that accurately in the first place. I mean, he was under no pressure for precision, and counting all those peas is incredibly dull and time consuming. I wonder how much of those numbers are down to estimates, rather than exact counting.

  13. nomdeplume says

    Nice one PZ, it had never occurred to me that using Mendel as a starting point (as I and all my contemporaries did) had that effect on our attitudes to genetics. I’m sure you are right although in my early research I quickly became aware of pleiotropy, and later became aware of genes which don’t have dominant/recesssive attributes. Looks like Biology should stop teaching with historic development as a template.

    @4 Rob – yes, religious beliefs are incompatible with science, do you think they are compatible?

  14. Rich Woods says

    @birgerjohannson #2:

    It is surprising development is robust enough to produce functioning adults!

    There, see! Evidence for a loving, designer God!

    (I’ll show myself out.)

  15. hemidactylus says

    There’s much to be said for the transmission vs expression genetics dichotomy and also the evolutionary vs molecular gene concepts. Mendelian stuff is transmission simplified. Dawkins’ “selfish gene” metaphor is the evolutionary concept simplified though I recall him planting a subverting seed with the rowboat analogy.

    The expression and molecular concepts are much messier. The former has stuff such as cell-cell interaction and epigenetic thingies like methylation making it extremely more complex than notions one gets from simplified renderings of transmission and evolution.

  16. chrislawson says


    Mendel didn’t count peas, he counted plants. It’s still a lot of counting.

  17. cvoinescu says

    dragon hunter @ #12, from what I remember from The Selfish Gene, there is a discussion about what is a “gene”, and Dawkins intentionally doesn’t settle on a definition. He just uses the word as a fuzzy shortcut to mean any stretch of DNA that has an effect over its own reproduction (in the context of all other “genes”, as he often insists and you rightly point out, and usually via the phenotype; but he also discusses transposons and repeating elements). I think he also touches on the fact that “phenotype” is not a clear-cut concept either; at one end of the scale, he later expands that into a whole book, The Extended Phenotype. So I think he covers your points in the book, but you’re right that the Reader’s Digest version of the concept misses most of these subtleties.

    Where The Selfish Gene lacks is the other end of the phenotype scale, at the molecular level. I don’t think there’s a clear demarcation, even: it’s hard to draw the line between what we now call “proteomics”, “metabolomics”, “transcriptomics” and even “epigenetics”, on one hand, and “genomics” and “phenomics” bracketing then at either side of what I see as a tangled continuum. Admittedly, these interactions were much less well understood at the time the book was written (and our understanding is still not great).

  18. cvoinescu says

    I don’t usually double-post, but when I do, it’s because I forgot something important. [takes sip]

    Two things, actually.

    Dawkins repeatedly points out that evolution must be understood in the context of populations (and thus, among other things, gene pools).

    The “selfish gene” idea (in its subtle and many-layered form, not the simplified caricature) is, in a way, the mechanism Darwin was so painfully aware that was missing from his theory. It is, in a way, the same idea, in generalized form. I seem to remember that Dawkins points out that the fact that genes (“genes”) are grouped together inside individuals makes that layer the most prominent unit of selection, and that level makes the same predictions as “classical” evolution. That that’s not the only level is what’s interesting, and the details (often mathematical) of how some of those situations happen are quite cool. For all Dawkins’s many later failings (and I do include some of his more recent books), The Selfish Gene is well written and explains the concept and some of its subtler aspects very well. I’m torn on whether the sound-bitey, click-baity name for the concept is a useful metaphor or a pernicious oversimplification, but I have no reservations regarding how it’s treated in the book. (Also, it’s the book that made the idea of meme popular; can’t really hold it responsible for how that turned out several decades later.)

  19. chrislawson says

    As for the implication that his figures are too good to be true: the biggest dataset in his paper is his trial of 639 hybrids, which ended up in the phenotype ratio of 10:19:43:78 against a mathematically derived ratio of 10:20:40:80, which in his own words “agree so closely…that this last undoubtedly represents the true value.” Yeah, that closeness looks like the sort of accuracy you can expect with millions of samples, not hundreds.

  20. chrislawson says


    I wonder how much better it would have been if Dawkins had called his book The Selfish Replicator.

    It still has the problem of attaching motive to inanimate entities, but at least it makes sense as a metaphor, that is, we can’t expect molecular-level replicators to display moral values.

  21. chrislawson says

    Correction: Mendel did count peas as well as plants. My apologies. I misread the paper by skimming to the data sets.

  22. Rob Grigjanis says

    nomdeplume @14: AFAIC, whether a particular scientist’s particular beliefs are compatible with their work is up to them to decide. And, of course, you’re free to believe in your little mantra, for whatever good you think it will do.

  23. says

    Did Mendel cheat? I don’t know. I think he sort of fudged. One interesting point from the paper I cited is that Weldon did examine the pea varieties that Mendel used, and for instance, the yellow vs green dichotomy he used wasn’t quite as crisp as Mendel would lead you to think — there was a gradation of colors. I think Mendel knew what answer he wanted and consciously or unconsciously biased his assessment.
    What he should have done is a blind test — recruited another monk to score the peas with no knowledge of the expected results. That wasn’t an option, I guess. All the work was done on his own. It’s a bit late to criticize the methodology.
    Was this a crippling flaw to his work? Maybe, but we don’t rely on Mendel’s data to interpret modern work, so it’s rather irrelevant. Should we ask Verhandlungen des naturforschenden Vereines im Brunn to issue a retraction?

  24. says

    I predict that the ghost of Gregor Mendel is going to ‘pea’ on Jordan Peterson and his super-microscopic DNA viewing tiny mind, said Sarcasimus Maximus.

  25. imback says

    When your education is arrested at that point, you start thinking in terms of “genes for X”, where X is cancer, or obesity, or eye color, or sexual preference.

    It seems to me that eye color is not like the others in this list. Cancer, obesity, and sexual preference depend on a multitude of genes as well as one’s prenatal and postnatal history. But eye color, while not so simple as being determined by a single locus of a dominant or recessive gene, can have some rough Mendelian-ish probabilities predicted for it just from knowing the eye colors of the parents. (This is from my own observation but maybe I’m wrong; I have not kept up with any research on this.)

  26. says

    Known genes that affect eye color: OCA2, HERC2, ASIP, IRF4, SLC24A4, SLC24A5, SLC45A2, TPCN2, TYR, and TYRP1. The Mendelian distribution is fairly approximate. Remember that eye color is affected not just by the pigments, but also by their patterning and the optical properties of the overlying tissues.

  27. Pierce R. Butler says

    PZ Myers @ # 25: … we don’t rely on Mendel’s data to interpret modern work…

    Kinda hard to avoid “interpreting modern work” which uses phrases like “meticulous accounting” in present context, except perhaps as submeticulous.

    Given that “present context” also includes serious pain and serious meds, I trust the Relevant Authorities will let this one go by with no more than a routine alert to the Adjective Police.