Higher order thinking

The one thing you must read today is David Dobbs' Die, Selfish Gene, Die. It's good to see genetic accommodation getting more attention, but I'm already seeing pushback from people who don't quite get the concept, and think it's some kind of Lamarckian heresy.

It's maybe a bit much to ask that the gene-centric view of evolution die; it's still useful. By comparison, for instance, it's a bit like Mendel and modern genetics (I'll avoid the overworked comparison of Newton and Einstein.) You need to understand simple Mendelian genetics — it gives you a foundation in the logic of inheritance, and teaches you a few basic rules. But once you start looking at real patterns of inheritance of most traits, you discover that it doesn't work. Very few traits work as Mendel described, and one serious concern is that we tend to select for genes to study that behave in comprehensible ways.

And every geneticist knows this. Mendel was shown to have got some things wrong within a decade of his rediscovery: Mendel's Law of Independent Assortment, for instance, simply does not hold for linked genes, and further, linkage turns out to hold important evolutionary implications. But I still teach about independent assortment in my genetics course. Why? Because you need to understand how to interpret deviations from the simple rules; it's an "a-ha!" moment when you comprehend how Morgan and Sturtevant saw the significance of departures from Mendel's laws.

Most of genetics seems to be about laying a foundation, and then breaking it to take a step beyond. Teaching it is a kind of torture, where you keep pushing the students to master some basic idea, and then once they've got it, you test them by showing them all the exceptions, and then announcing, "But hey! Here's this cool explanation that tells you what you know is wrong, but there's some really great and powerful ideas beyond that."

That's what's fun about genetics: compounding a series of revelations until the students' brains break, usually right around the end of the term. Over the years I've learned, too. Undergraduate genetics students usually collapse in defeat once I introduce epistatic interactions -- the idea that the phenotype produced by an allele at one locus is dependent on the alleles present at other loci -- but it's always great to see the few students who fully grasp the idea and see how powerful it is (future developmental biologists identified!).

And that's how I see the gene-centric view: absolutely essential. You must understand Mendel, and Fisher, and Wright, and Hamilton, and Williams, and once you've mastered that toolkit, you can start looking at the real world and seeing all the cases where it's deficient, and develop new tools that let you see deeper. The new idea that Dobb's describes, and that is actually fairly popular with many developmental biologists, is that phenotype comes first: that organisms are fairly plastic in response to the environment in ways that can't be simplified to pure genetic determinism, and that the genes lag behind, acting to consolidate and make more robust adaptive responses.

I've written about genetic assimilation/accommodation before, and have given one lovely example of phenotypic change occurring faster than the generation of new mutants can explain. It's always baffled me about the response to those ideas: most people resist, and try to reduce them to good old familiar genes. It's a bit like watching students wrestle with epistatic modulation of gene expression when all they understand is Mendel, and rather than try to grasp a different way of looking at the problem, they instead invent clouds of simple Mendelian factors that bring in multi-step discrete variations. They can make the evidence fit the theory -- just add more epicycles!

I'm seeing the same responses to Dobbs' article -- it's still all just genes at the bottom of it, ain't it? Oh, sure, but the interesting parts are the interactions, not the subunits. We need to take the next step and build tools to study networks of genes, rather than reducing everything to the genes themselves.

More like this

This article further confirms in my mind that genetic determinism is the new Behaviourism, what Arthur Koestler might have called 'Flat Earth biology'.

The Dobb's article is spot-on.

There are "selfish genes". Homing endonuclease genes (HEGs) are extremely selfish and parasitic.

I strongly suspect that it is the existence of HEGs that limits the fidelity of eukaryotic DNA replication. There has to be enough copying error for HEGs to not persist indefinitely or that species goes extinct.

Copying error isn't a “bug”, it is a “feature”.

By daedalus2u (not verified) on 03 Dec 2013 #permalink

One problem with this beautifully written article (and with everything it's based on, is the failure to distinguish between two meanings of the word 'gene'. First, 'gene' is commonly used to mean 'locus' - a stretch of sequence with a specific function, usually coding for a specific protein. Second, 'gene' is also commonly used to mean 'allele', a version of a sequence or locus.

Geneticists typically slip back and forth between these two meanings, and in most contexts that's OK. But in a discussion about the importance of 'genes', carefully distinguishing between the two meanings is essential.

By Rosie Redfield (not verified) on 03 Dec 2013 #permalink

I'm just an amateur evo-devo enthusiast, and have trouble following the technical details once they get beyond pop-sci level. But I don't see why any of this makes Dawkins' idea of "The Selfish Gene" wrong. Whether a gene is (simplistically) a stretch of DNA that codes for a specific trait, or a combination of complex factors that result in changes to the phenotype through various means, his metaphor still applies: all they "do" is compete to be repeated and preserved in future offspring. To show his "selfish gene" metaphor to be wrong, you'd have to come up with another driving force for evolution besides natural selection of genetic traits. Right?

By Peter Sagal (not verified) on 04 Dec 2013 #permalink

Peter, it's not (as I understand it) that the genes aren't selfish, it's that they don't serve as the full explanatory basis for phenotypical features. This is why a cloned animal can look very different from its "parent," even though it started out with the same genes.

By Chris Cogan (not verified) on 04 Dec 2013 #permalink


Please note that I am being fair to you about this under-
statement you made in this post . . .
Even though I'm trying to become a geophysicist, and I
was telling you the truth on the phone about Doctor Bruce
Buffett, Chairman of Berkeley Geoscience, telling me that
"People don't have a good handle of what caused the Sup-
ercontinent to break up. We don't know what the energy
source was," and that you can't name one college or lab
which has delved into all the water Earth could have had under all the Supercontinent it could have had, you were
misrepresenting that in your 5 September post about our
phone conversation 4 Sept. My presupposition you didn't
ation with measurement. Do you know (or does any geo-
physicist know [Maybe the dude who commented I AM a geophysicist?] how powerful our planet's inner lightnings were when it went from not having to having its Mid-Oce-
Can you call Dr. Walt Brown and point out to him what
you wrote about in "Happy Atheist" was doing justice to
his www.creationscience.com ? It says nothing about a
combustion of H and O in the atmosphere (Strawman).
Please note that I majored in molecular biology with my
earth science as part of my 129 undergraduate credits &
I hope that next time you post a blog about me you have
me write out an e-mail to you knowing beforehand that it
will be posted. Your "--looking for a geophysicist" post on
me received comments as if it was my "best foot forward".
We will try an e-mail exchange involving you posting it la-
So here's one that's better prepared about your under-
statement: "the phenotype produced by an allele at one
locus is dependent on the alleles present at other loci"
NOW, I don't play like that is all ya got! You can write
better on to successfully carrying research through to practical applications for genetic medicine. You know
the prerequisite is a full elucidation of all of our genes, including their expression patterns in distinct cell types,
how genetic expression and repression patterns in distinct cell types effect the anatomy and physiology of each cell's form and regulatory components from their initial divergence after the eight (divided) cell stage from a fertilized egg (differentiating the roles they begin to play; the production of specialized cell types; change [maturation] of cells through programs that turn on and off genes at certain times), alternative chromatin (complexes of DNA / proteins) states, correlation between chromosomal localization and coding expression patterns, sequences with which gene products appear or disappear, locations of specific gene products at specific development stages; direct molecular interactions, combinational or hierarchical control of signaling, which genes are turned on and off for the various differentiation pathways, the functional class of each gene, changes elicited in specific cells from their previous histories, e.g., effects of earlier signal transduction pathways, which cells have genes primed to respond to activity regulating transcription factors, which cells have genes made inaccessible to regulation, which programs have certain signal transduction pathways (involving biomolecules that carry signals between cells, within cells, and interpret those signals) regulating cells' transcriptional activities--how the properties of cells are changed by developmental signals (some of which occur to pathway components modified by specific content and activity of enzymes or change in location of cell-surface receptors either laterally or internalized subsequent to signal transmitting molecule binding independent of gene activity modification; integration of a barrage of overlapping relationships between signal transduction pathways and other regulatory molecules, organization of functions around irregularities that are set up by external events such as fertilization, gravity or its graded distribution of chemicals that the mother incorporates into an egg cell [asymetrical disribution of chemcials within a cell before it divides, and partitioning afterwards, sending two sibling cells to different fates], induction [the abilities of one group of cells to influence developing of another]; migration of cells both individually and adhered to other cells; selective association of cells involving changes in substances on the surfaces of cells [cell adhesion molecules] which vary either in amount or chemical identity from one type of cell to another, and facilitate or disrupt when and where cells adhere to each other, extension / contraction properties of cells involved in the various morphogenic movement shaping an embryo, lateral inhibition [when the touching of cell surfaces inhibit movement; and in a turning off of genes that function in cell division, reproduction and controlling death* of cells).

Keep researching molecular biology!
Cheers, Rick

By Rick Keane (not verified) on 05 Dec 2013 #permalink

ation with measurement.

What nonsense. Things of similar ages – the moon, carbonaceous chondrites, and so on – contain very similar amounts of heavy radioactive elements, as we would expect from the way supernovas work.

Do you know (or does any geo-
physicist know [Maybe the dude who commented I AM a geophysicist?] how powerful our planet’s inner lightnings were

There are no lightnings in solid rock.

when it went from not having to having its Mid-Oce-

...You seem to believe that there was no plate tectonics before Pangea broke up. This has been known to be wrong for half a century. It's in fact quite obvious: where did you believe the Appalachians or the Caledonides come from? The continents have been bumping into each other like go-carts for at least two and a half billion years, most likely for 4.4 or more. The Earth hasn't always had the mid-Atlantic ridge or the Atlantic ocean (both are some 160 million years old); it hasn't always had the East Pacific ridge (some 190 Ma old) or the Pacific ocean (some 750 million years old, though no ocean floor from that time remains); but it has had midocean ridges for most, likely almost all, of its history.

Finally, your last sentence takes up an entire screen. It would be quite laughable to believe that anyone is going to read it.

By David Marjanović (not verified) on 05 Dec 2013 #permalink

@PZ Myers: I don't know how you manage to keep your cool with those word salad commenters!
Meanwhile 'Inner Lightnings' does seem like a 'Earth Wind Fire' type band name throu?

@PZ Myers: I don’t know how you manage to keep your cool with those word salad commenters!

He's used to it, he doesn't have time to read every comment, and... he hardly ever reads the comments on this version of Pharyngula. The complete version is over on http://freethoughtblogs.com/pharyngula/ .

By David Marjanović (not verified) on 06 Dec 2013 #permalink