Kostas Kampourakis is a respected scientist who is being promoted by the National Center for Science Education (NCSE) as an excellent communicator of evolution. Here's a short video (below) where he explains why teachers are making a mistake by saying that Mendel is the father of genetics and that simple Mendelian genetics can explain complex traits.
I'm struggling to understand his point. Here's what I think he means.
I believe that the fundamental trait is the lack of an enzyme for degrading chlorophyll and this is what I would teach my students. I would also show them that the phenotype can be easily explained once you understand the biochemistry. It shows you that the connection between the fundamental trait and the visible phenotype can be mysterious so you should be careful about jumping to conclusions.
I think that Kampourakis sees this differently. He thinks that the example of green vs yellow peas is used to teach students that the color of peas is completely determined by a single gene. He thinks that the "trait" is the develpment of seed color in peas.
Kampourakis believes that "people looking for explanations and whatever happens to them in terms of disease and their own features, they find hard to reconcile the simplistic model that they have been taught with the realities of life." I think what he means is that students are being taught that single genes will always determine complex characteristics. He attributes that to the teaching of Mendelian genetics.
If he is correct, then that kind of teaching has to stop but I don't think it's the fault of Mendelian genetics. Mendelian genetics—indeed the entire field of genetics sensu stricto—is about the segregation of alleles. It is not about development even though we have come to learn a lot about development through genetics and the phenotype of mutants. I think that genetics and development are separate topics.
Kampourakis disagrees. He says, "we need, I think, to teach genetics from a developmental perspective. We need to show that genes do not determine traits but they are implicated in development." I believe this perspective comes from a more fundamental bias that distinguishes his worldview from mine. I tend to see genetics as a subject that covers all of biology and that includes all species such as bacteria, viruses, and single-cell eukaryotes. He tends to see things from a human perspective, which is much more complex than dealing with simple organisms. I think we should concentrate on teaching students about simple well-understood model organisms and then move on to explaining how this applies to more complex organisms. Kampourakis seems to be implying that we should jump right into reaching high school students about the most challenging issues in biology.
1. Actually, the common allele for white eye in Drosophila (see image) is X-linked so it doesn't follow the standard rules for Mendelian segregation!
15 comments :
I suppose we could teach intro to economics without discussing microeconomics....
With you on this, as usual
The idea that there is one gene for eye color, one for earlobe shape, one for this, one for that died in about 1920 as geneticists found more and more examples of pleiotropy. Over the next decade or two quantitative geneticists came to understand that a simple model of determination of traits by adding up effects of alleles was a very useful approximation, even though of course these loci did interact too. In 1917 Sewall Wright published a series of papers reviewing coat color genetics in multiple species, and suggesting that the genes acted by producing enzymes, and that one could explain the interactions by the action of these enzymes in the coal pigment pathways. (I once asked Sewall Wright if he had been the first to suggest that genes produced enzymes, and he said oh no, Cuenot had suggested that in 1904). So Kampurakis is a bit out of date.
Oops: "coal pigment" --> coat color pigment
"Kampurakis" --> Kampourakis
William Bateson began correspondence with the physician Archibald Garrod in January 1902. The latter developed the one-gene-one-enzyme idea in his 1909 text Inborn Errors of Metabolism. In 1934 Garrod noted that "it was Bateson who saw the daylight". For details see p. 195 of "Treasure Your Exceptions." The Science and Life of William Bateson (2nd edition 2022).
It is also worth noting that a common kind of genetic experiment was to make two inbred lines that differed only at one locus (or at two possibly-linked loci). Knowing how Mendelian genetics works is essential to understand the results of crosses in those cases.
Before Crispr/cas9 was available the need to backcross mouse mutants for 10 generations to end up in a pure genetic background to reduce the effects of other genes. For the very same reasons the Jackson lab does quite some backcrossing to reduce effects cause by genetic drift.
But I'm in favor of the cell -centric view of life not gene-center perspective --supposedly I assume that fellow is emphasising on that view which is more holistic and complete
Mendelian genetics—indeed the entire field of genetics sensu stricto—is about the segregation of alleles. It is not about development even though we have come to learn a lot about development through genetics and the phenotype of mutants. I think that genetics and development are separate topics.
I completely disagree with this. Genetics as a field came about through through studying the inheritance of physical traits (the phenotype) and explaining how they are passed on. The phenotype was the key starting concept, and the observation that the phenotype was inherited (or sometimes not) across generations was the very subject of genetics from the get-go. There would be not genetics without the phenotype. The only exception I can think of is "pure" population genetics regarding situations were you are only concerned with alleles and their frequency (nothing said about phenotypes in this case). However, that's a far from generalising the entire field of genetics (etiam in stricto sensu). This is especially true regarding Mendelian genetics, the first law of which is the law of dominance, which does not make sense without the phenotype.
So, both genotype and phenotype are key concepts in the study of genetics. The inheritance of the phenotype was the observation to be explained. The passing on of the genotype (alleles) was the explanation. However, the concept of the phenotype is incomplete without ontogeny. The phenotype is produced through ontogeny. One could argue that ontogeny is (part of) the phenotype. Without ontogeny, the connection between the genotype and phenotype becomes a blackbox, and you end up in a situation akin to that classical cartoon "then a miracle occurs".
@Nesslig20: Is this just an argument about names? About what is called "genetics". It seems that it is not about the biology of anything.
@Joe Felsenstein
No. This is about the facts of history. The inheritance of the phenotype was the phenomenon to be explained, and those interested in answering this question started the field of genetics. However, the inheritance of the phenotype is incomplete without ontogeny. Genetics without development doesn't explain the phenotype, yet it is still commonly believed through notions such as "the genome is the blueprint for the organism".
@Nesslig20 The phenotype is simply a marker for an allele. Genetics explains the segregation of alleles. Biochemistry, molecular biology, and other fields explain the connection between the allele and the phenotype.
Some people think that the field of genetics includes all of the things that explain phenotypes, including transcription and translation. Are you one of those people?
@Larry Moran
The phenotype is simply a marker for an allele.
Sorry, but that's a terrible definition. That's even worse than what I was taught in high school. If we adhered to this definition, then whenever we encounter a distinctive phenotype we would assume that there is 'an allele' (singular) corresponding to it. Of course, this is wrong in almost every case you can study in real life. Even Mendelian genetics gets this (more) correct. It at least takes dominance/recessiveness into account; in which case the phenotype isn't a "marker" for the allele, since he expression of the corresponding phenotype can be prevented by a dominant allele (and in non-Mendalian genetics via epistasis).
Alright... Now, to be charitable, I assume you meant to say that “the phenotype is simply a marker for the genotype” and the genotype is the complete set of alleles (the entire genome) which accounts for dominance/recessive effects, as well as epistasis and pleiotropy, etc. But still, we all understand that, even from identical genotypes, you won’t necessarily end up with identical phenotypes. Drastic examples include the different castes (workers and queens) in eusocial insects. In this case, the phenotype (queen) is a “marker” of non-genetic stimuli (royal jelly). But here we run into the danger of the ‘nature-nurture’ canard, since that’s also not quite right.
Even if you account for all the genetic and environmental variables, you still haven’t explained the phenotype. It still commits the error of trying to rescue a form of preformationism. Somehow the “plan” for the organism was there from the beginning (predetermined), and we can find it if we searched more in the genome (or in combined genome+environment).
What actually happens is a state space with bifurcations and attractors in ontogeny, akin to Waddington’s concept of the epigenetic landscape. A developmental trajectory may tend towards one attractor by default (worker) but one change can alter the starting position of the trajectory, or it can change the landscape itself all together, such that the trajectory now tends towards a different attractor (queen). And sometimes, you don’t have stable attractors but very highly variable states that are at the mercy of noise (e.g. the development of finger prints). Thus, the phenotype is not a “marker” of anything. It’s a process in and of itself, which continues overtime.
For those who are interested to dive more into this, I highly recommend Johannes Jaeger’s papers and talks. He is a good example that applies to this topic:
* Genetic Causation in Complex Regulatory Systems: An Integrative Dynamic Perspective
Also a good video on his youtube channel:
* 07.04 Feedback, Levels, and Causation – Beyond Networks: The Evolution of Living Systems
Some people think that the field of genetics includes all of the things that explain phenotypes, including transcription and translation. Are you one of those people?
I mean... I was taught, and I have taught students, what transcription and translation are in courses on genetics. They are central topics in molecular genetics. So, yes. Feel free to shoehorn me in as one of "those people".
@Nesslig20 says, "I mean... I was taught, and I have taught students, what transcription and translation are in courses on genetics. They are central topics in molecular genetics. So, yes. Feel free to shoehorn me in as one of "those people"."
Thanks. That explains a lot. We are talking past each other. I think of "genetics" as simply the field that studies the segregation of alleles. It includes population genetics. (Yes, this is a bit of over-simplification but I think you get the idea.)
You tend to think of "genetics" as a much broader field that includes the study of what genes do and how they do it. That's why you are so interested in explaining the role of genes (and other parts of the genome) in producing visible phenotypes.
So, when I said that a phenotype (e.g. smooth vs wrinkled peas) is simply a marker for an allele, I meant that the fundamental problem was to determine the principles of allele segregation before we knew about DNA sequences.
You replied, "If we adhered to this definition, then whenever we encounter a distinctive phenotype we would assume that there is 'an allele' (singular) corresponding to it." That's because you want to explain phenotypes, not allele segregation. I never meant to imply that every heritable phenotype has to be explained by the segregation of a single allele. I only meant to say that there are times when we can study the fundamentals of genetics by using particular phenotypes as markers for the segregation of single pairs of alleles.
@Larry Moran
”That's because you want to explain phenotypes, not allele segregation.”
No, I was using your definition and critiqued its implications.
”I never meant to imply that every heritable phenotype has to be explained by the segregation of a single allele”
I already assumed this to be the case (see second paragraph of previous reply).
”I only meant to say that there are times when we can study the fundamentals of genetics by using particular phenotypes as markers for the segregation of single pairs of alleles.”
That’s very different than saying “The phenotype is simply a marker for an allele”, but I digress. As both you and I are aware, the times when one even can use phenotypic variation to follow the segregation of alleles (especially single pairs) are VERY few and far between. The vast majority of phenotypes are inherited in a non-mendelian fashion. Mendel was very lucky and even he know that his laws only applied to certain ‘variable hybrids’, not across the board as his experiments on Hieracium (Hawkweed) showed.
“We are talking past each other. I think of "genetics" as simply the field that studies the segregation of alleles. It includes population genetics. (Yes, this is a bit of over-simplification but I think you get the idea.).
I don’t think it is an oversimplification as much as it is a far too narrow (and flipped) view of ‘genetics’. As I have explained before, the inheritance of the phenotype across generation was the phenomena to be explained. We didn’t know about genes or alleles, but we knew that children looked like their parents. The questions of “why” and “how” was the birth of genetics as a science. But today, you define ‘genetics’ as the segregation of the alleles and the dynamics of allele frequency’s in populations. This pushes the phenotype to the background, no longer the central subject of inquiry. Now the phenotype and it’s inheritance doesn’t need any explaining. Not by geneticists at least. Figuring that out has become someone else’s job. Geneticists may use the phenotype as a marker to follow the segregation alleles if the phenotype happens to be correlated to an allelic pair (which is rare). The reason why I called this ‘flipped’ is because the inheritance of the phenotype was the phenomenon to be explained, and the “allele” was conceived as inheritable factor whose physical nature was not known at first. Now the allele is at the center and the phenotype orbits around it.
I am not saying you can’t do that. Anyone can come up with their own definition of anything. I am just saying that this stands in stark contrast to history. Lastly, this is also moot when it comes to discussing the issue that Kampourakis brings up. His concern is that teaching Mendelian inheritance to students may lead them to the cognitive trap of genetic determinism. How can we avoid this? Just telling them not to worry about it, since we define genetics simply as the field that studies the segregation of alleles? I don’t see how that helps.
Post a Comment