Wednesday, May 03, 2017

Debating philosophers: The difference between genes and alleles

This is my third post on the Lu and Bourrat (2017) paper [Debating philosophers: The Lu and Bourrat paper]. Part of their argument is to establish that modern evolutionary theory is a gene-centric theory. They need to make this connection because they are about to re-define the word "gene" in order to accommodate epigenetics.

In my last post I referred to their defense of the Modern Synthesis and quoted them as saying that the major tenets of the Modern Synthesis (MS) are still the basis of modern evolutionary theory. They go on to say,
One of these tenets, which will be the focus of this paper, is that phenotypic evolution can be explained by changes in gene frequencies in a given environment. This "gene-centric view," relies on genes being the sole heritable material, which, together with the environment, determine the phenotype.
Evolutionary theory has been based on population genetics since the 1920s and 1930s. That's almost 100 years. It's only in that sense that evolutionary theory is "gene-centric." It's more appropriate to refer to it as "genetic."

Calling evolutionary theory "gene-centric" seems to come from Richard Dawkins and his 1976 book The Selfish Gene. I recently pointed out that his view of evolution is not the dominant view among today's evolutionary biologists [The selfish gene vs the lucky allele].

In their discussion of standard evolutionary theory (SET), Lu and Bourrat refer to proponents of a new extended evolutionary synthesis (EES) for their information on what the current view is. They say,
SET, which EES proponents believe retains the core of the MS, has the following three tenets: ‘new variation arises through random genetic mutation; inheritance occurs through DNA; and natural selection is the sole cause of adaptation, the process by which organisms become well-suited to their environment’ (Laland et al. [2014], p. 162).
Now, nobody doubts that variation arises through mutation and that inheritance involves DNA, although very few others would express it that way. The idea that natural selection is the only means of adaptation is debatable but that's not so important. What's important is that there's no mention of non-adaptive evolution and that's a very important part of standard evolution theory in 2017. Furthermore, I would argue that changes in allele frequency in populations is an important part of evolutionary theory.

My major criticism of Lu and Bourrat's paper has to do with their linkage of evolutionary theory and "genes." This is an immportant part of their paper and there will be two other blog posts devoted to their description of the "evolutionary gene" and the "molecular gene." They make their task much more difficult by confusing genes with alleles. This is the same problem I wrote about a few days ago [The selfish gene vs the lucky allele].

Here's an example of the confusion,
Gene selectionism represents a strong version of the gene-centric view of formal evolutionary theory (Hull [2000], p. 422; Laland [2004]). Haig ([2012]) develops the notion of the ‘strategic gene’ in accordance with the common characterization of evolution as ‘changes in gene frequency and phenotypic effects of these changes’. For him, a gene refers to a determinant of difference in the phenotype that correspond to a set of gene tokens, mainly DNA pieces. The crucial point we retain from Haig’s account is that a gene in an evolutionary context is a difference maker. For defending gene selectionism, Haig ([2012], p. 470) regards a gene as ‘a strategist in an evolutionary game played with other strategic genes’, hence his use of the term ‘strategic’.
I don't see any good reason for not using "allele" in the paragraph above. It would be much more accurate, especially when describing the definition of evolution as, "changes in gene frequency."

There are two problems with "gene:"
  1. It creates an unnecessary conflict between the definition of the evolutionary gene and the molecular gene. It's not really "gene" frequencies that change; it's allele frequencies where alleles are gene variants that are created by mutation.
  2. Not all heritable changes are found in genes but they can all be classified as alleles. For example, variants in regulatory regions are arguably not part of genes but are very important in evolution. Variants in DNA sequences at the sites of centromeres, telomeres, SARs, origins of replication, and even junk DNA can be relevant in evolution and these are not "genes" in any reasonable sense of the word.
There's no reason for these philosophers to ignore the standard use of the word "allele" as in "changes in allele frequency." It's an example of fuzzy thinking. So is referring to current evolutionary theory as "gene-centric."

Image Credit: Campbell Biology


  1. One more time - It is very depressing how much of this discussion and commentary focuses on the low standards of 'amateurs' and 'autodidacts' and 'philosophers'. This is an argument from authority. It is also a bit silly, as both these authors have biology degrees, and one has published a number of articles in behavioral ecology in their previous career as a scientist. I say that just to neutralise the argument from authority. Now let's focus on the substance!

    1. It would be an argument from authority were it not because the criticism is followed by explanations. There's substance. More substance can be found in prior posts. I don't always agree with Larry. But I have never caught him giving no explanations for his criticisms (except when he has given those explanations many times in previous posts).

    2. Paul,

      I don't understand why you're saying this is an argument from authority.

      My criticism of Lu and Bourrat is that they are not describing current evolutionary theory even though they claim (implicitly) that this is their area of expertise.

      Part of what I am doing is pointing out that there are different "authorities" with very different views of modern evolutionary theory and very different views of the important distinction between "genes" and "alleles."

      Of course there's a bigger picture that's related to the discipline of philosophy. In my experience, philosophers of biology often publish articles that make no sense to the scientists in the discipline. This is one of those papers.

      In my next post ('The Molecular Gene") I'll be discussing the definition of a molecular gene. Lu & Bourrat use the definition you and Karola promoted in your 2013 book.

      By any reasonable definition you are an "authority" on the definition of "gene" and that's why Lu & Bourrat reference you. I will show that your definition is not accurate and is not the one that should be used in molecular biology.

      I will be attacking an authority in the philosophy of science; namely, you! (Sorry!) I look forward to your comments. :-)

    3. Larry writes: "In my experience, philosophers of biology often publish articles that make no sense to the scientists in the discipline [i.e. to biologists].

      This suggests the question: To whom should an article on Philosophy of Biology make sense?
      a. Philosophers.
      b. Biologists.
      c. Both, otherwise what is the point?
      d. The authors only. Everyone else has to deconstruct the text to generate meaning.

    4. Didn't mean to imply there is not substance here *as well*. But I am a bit sensitized to 'philosophers versus biologists' since a few years ago a fellow philosopher dismissed my views as 'but perhaps biologists are right' when citing work in which the silly 'philosopher' view was me collaborating with Fred Nijhout, Pat Bateson, etc 😂

  2. I have to shamefacedly admit that the confusion between "gene" and "allele" was originally bad terminology perpetrated by population geneticists. The frequencies are of course frequencies of alleles, but they are routinely called "gene frequencies".

    The whole question of what "a gene" is depends on the context, in an unfortunate way. If we are members of a population of 1,000 individuals, how many "genes" are there? 20,000? 20,000 x 2? 20,000 x 2 x 1,000? At one particular locus, say my malate dehydrogenase locus, do I have one gene? Two genes? Or as many genes as I have different alleles (1 or 2)?

    We're not muddled -- we just have bad terminology. In each case where we use the word, we can figure this out from context. But we shouldn't have to.

    And this is aside from the issue of how far out from the start and end of the coding sequence the "gene" extends.

    1. The human genome has about 20,000 protein-coding genes and about 5,000 genes for noncoding RNAs. The number of genes for noncoding RNAs is in dispute. We are certain there are at least a few thousand.

      The average diploid cell has 50,000 genes but to be precise you should say there are two copies of each of the 25,000 genes.

      There are many examples of gene families and in some cases the different members of a gene family in the same genome are identical. Nevertheless we say there are two (or more) different genes because they are at distinct loci in the chromosome. That's why each of the histone 2A genes, for example, has a different name in the gene databases (e.g. HIST1H2BA and HIST1H2BE).

      I believe that at any one locus you have one gene but it can exist in several different variants called "alleles." You may be homozygous at that locus or heterozygous. There may be no alleles, two alleles, or dozens of alleles for each gene in the population. There's no confusion, in my opinion, if we are careful to distinguish between alleles and genes.

      A population of 1000 human individuals usually contains 25,000 different genes at two copies each per individual. Thus, there are 50 million genes altogether but this isn't a very meaningful number.

      In addition, we should not forget there are segregating variants in the population that are not part of a gene by any reasonable definition of the word "gene." I prefer to call these alleles if there are different versions at the same locus but some people only use "alleles" to refer to gene variants. They don't have a word to describe segregating variants at an origin of replication or a regulatory site.

      All of this is just my opinion, of course. I am not an authority on genes and genetics.

    2. Thanks for the more accurate figure, but I was mostly trying to point out the diverse ways the word "gene" is used. For example, we talk about a tree of genealogical relationship between copies of a single "gene" as a "gene tree". If there are 20 tips, does that not mean 20 "genes", even though all are copies for the same locus?

      For that matter "gene tree" is also used for something very different, a tree where the forks are gene duplication events. Now each tip is a different gene, a different locus.

      And, as you note, we can even talk about the "gene frequency" of the two alleles of a SNP, where the SNP is off in the middle of nowehere and not near any protein-coding locus or any RNA-coding locus or any regulatory site.

    3. It's a confusion extended by Dawkins's snappy 'selfish gene'. Really, it's a selfish allele. And it's as long as a 'reasonable' amount of recombination makes it. I've tried using the term 'evolutionary allele' for such units. Would that be inappropriate?

  3. "Allele" is really bad terminology - as I described in my 2011 "Ditching alleles" essay. People can't spell it, they can't pronounce it and many don't know what it means. "Gene" ate "allele"'s lunch. Gene is shorter and works better as a prefix (think genetic, geneology) it recombines better and has produced mutant offspring (think meme, memetics). It's a much fitter term. 'Gene locus' is usable in the rare cases where one wants to talk about what is happening at a particular location. I don't think there's any excuse for promoting the "allele" terminology these days. It is just a really shitty word. People should use "gene" instead - and most folk have evidently been doing exactly that for a long time now.

    1. But "different alleles" and "different genes" mean quite different things. How would you deal with that?

    2. There's a question I've wanted to ask you John. Can an allele be made up of more than one gene?

    3. "Gene locus" or "genetic locus" is my preferred terminology for talking about a set of gene instances which act as functional substitutes due to their position.

      I hardly ever use these terms, though. Usually "gene" is a appropriate. "Dominant gene, recessive gene, gene frequences, ancestral genes, gene for X - and so on - they all work better if the term "gene" refers to a genetic instance rather than a genetic locus.

      It's similar to what happens in English - where the term "word" is used a lot, while the terms like "adjective" and "adverb" (which refer to sets of words that can substitute for each other) are used less frequently.

      I do have to explain my usage sometimes - which *is* an inconvenience - but I can hardly support passing this appalling terminology on to countless future generations. "Allele" is destined for the trash bin of history. IMO, we should all be doing our bit to hasten its demise.

    4. "Gene locus" or "genetic locus" is my preferred terminology for talking about a set of gene instances which act as functional substitutes due to their position.

      So you would use "gene locus" instead of "gene" and "gene" instead of "allele". Something tells me this is not going to clear things up. The problem is not in the terminonology, but in (as you said) that many don't know what it means.
      The same is happening now for example with the word "SNP". Most of the times, when I hear people talk about SNPs, they refer to SNP alleles, not to the polymorphisms. Of course, you can invent a new terminology to clear that up, but it's just fighting symptoms.

    5. The reason why many people don't know what the allele terminology means is because is unpopular. The reason why the allele terminology is unpopular is because it sucks. I explained the main reasons for its suckiness in my previous comment.

    6. If you say so. I still find it a useful word. BTW, here is another term that will perish for the same reasons:
      epistasis => gene-gene-interaction

  4. Richard Dawkins in the second edition of the Ancestor's tale (p 53-54) discusses how allele and gene terms are used differently among biologists of different disciplines. He says that behavioral ecologists use gene term indistinctly to alleles whereas molecular biologists refer to alleles as variants of genes. I am not sure that the distinction is so clearcut among branches of study.