A phenotypic view of evolution Evolution in Structured Populations


Why genotypes are not just the sum of the genes

I want to continue with my theme that the patterning node is not just genotype, and that the genotype is not just genes.  In particular, I want to talk about gene interaction, and why the genotype is not the sum of the genes.  This is an area in which I am an active researcher (in fact as I write this I have a computer churning away on a rather large analysis), which can be a problem.  When you are too close to a subject often it is difficult to write anything that anybody but you can understand.  I will try to keep this general, but if I fail let me know and I will do what I can to clean things up!

First, off it is obvious that genes interact.  Genes themselves do not interact (don’t argue with me about modern molecular findings, I am taking a last millennium view of the gene).  Rather, genes produce products such as enzymes that work in enzymatic pathways.  This means that the gene products inevitably interact.  In a standard enzymatic pathway you might find that some chemical, Chemical X (with apologies to the Powerpuff Girls), is acted on by an enzyme, enzyme A, producing a product, chemical Y, which is then acted on by enzyme B to produce chemical Z.  (for some very obscure reason this image is not going into the post well.  click on it if you need to see it better.)

Screen Shot 2013-05-10 at 11.01.21 AM

The rate at which enzyme B converts Y into Z will depend to a large extent on the relative amount of Y and Z in the system.  If enzyme A is very efficient there will be lots of Y available, and enzyme B will work quickly.  If different individuals have variants of B that work at different rates these rate differences will show up in the rate at which Z is produced.  If, on the other hand enzyme A is very inefficient the rate at which B works will be a function of how much of chemical Y is available.  Variation in enzyme B will not be expressed since even very inefficient forms of B are using up Y as quickly as it is made.  This type of interaction, that is an interaction between two loci that affects the phenotype, is termed epistasis.

There are lots of examples of epistasis.  One of my favorites is coat color in Labrador retrievers.  In Labradors, dogs with a BB or Bb genotype are black, whereas those with a bb genotype are chocolate; that is, unless they have the ee genotype at a second locus.  Regardless of the genotype at the B locus ee dogs are yellow.


So, do we say that the B locus codes for coat color or that it has no effect?  The answer depends on the genotype at a second locus.  This is the crux of the biscuit (again apologies to Frank Zappa).  The effect of an allele on the phenotype depends on the background it is found in.

There are lots of potential interactions among two loci with two alleles, and even more with additional alleles or loci.  In the Labrador example we can turn the colors into numbers by assigning a 2 to black, a 1 to chocolate, and a 0 to yellow.  Then the table becomes:

dog epistasis numbers

The numbers are values for the phenotypes (often paradoxically called genotypic values), and in principle, for any pair of loci with two alleles we can fill in the appropriate phenotypes.

What I have just outlined has been called “physiological epistasis”.  Physiological epistasis can be thought of as the interaction among loci measured on an absolute scale.  Using physiological epistatic measures it is clear that the B and E loci interact epistatically.  It is important to contrast this with “statistical epistasis” which is whether or not there are epistatic interactions in a population.  In the above example, if the population was fixed for the EE (or ee) genotype we would say there were no epistatic interactions.  This is because only the first column would be observed in the population, and we would say that the B locus was a simple dominant locus.  Statistical epistasis is only seen when both alleles at both loci are present.  From a more nuanced perspective the amount of epistasis will depend on the gene frequencies at both loci.  In general the amount of epistasis will be greatest when both loci have a gene frequency of 0.5, and it will diminish as frequencies diverge from that intermediate value.

It is useful to consider an analogy to height here.  I am 5 feet 5 inches tall, and this is true whether I am measured in the Netherlands or Guatemala.  That is, my height is a fixed attribute that is independent of context.  Although my height is fixed, whether I am tall or short is not fixed, and depends upon where the comparison is done.  In the Netherlands the average male is about 6 feet 1 inch tall, and as a consequence I would be considered short.  In contrast in Guatemala the average male is 5 feet 2 inches tall, and I would be considered tall.  So, am I tall or am I short?  It depends on where I am being measured. (height data from http://www.disabled-world.com/artman/publish/height-chart.shtml)

Just as it is not possible to make statements about being tall or short outside of the context of the groups being compared, it is not possible to make statements about the effect of a gene on the phenotype independent of the genetic background in which it is measured.  If we like black dogs, and we are using a population fixed for the EE genotype we can say that the B gene is a “good gene” because it causes dogs to be black.  However, if we are using a population fixed for the ee genotype the B gene has no effect on the phenotype, and it would be considered a neutral gene.

The Elements of the Phenotype

In the next several weeks I will expand on my thoughts about how I think about evolution.  This week I will discuss my thoughts on the formation of the phenotype.   In the following weeks I will talk about the forces of evolution and how these should be generalized away from the gene-centric view that currently dominates evolutionary biology.

I find it useful to think of the phenotype being determined by what might be called a node of patterning interactions (patterning node for short), and non-heritable information nodes.  The patterning node is the product of a set of heritable information nodes.  I call these “nodes” in analogy with network theory.  That is, we can think of the heritable nodes “determining” the patterning node, which in conjunction with the non-heritable nodes determines the phenotypic node.  This is, of course, very similar to the traditional portioning of the phenotype into genetic and environmental components; however, this traditional partitioning is simplistic for at least two reasons.

Screen shot 2013-05-03 at 10.21.03 PM

First, there is a tendency to use a short hand of describing a “gene for something”, such as a gene for flower color, or a gene for altruism.  There is also a tendency to think of the genotype as the sum of the genes, and that the phenotype is the genotype as seen through a filter of the environment.  Both of these views miss the point that although the genes do represent transmitted information, the effect of that information on the phenotype is the result of a complex system of interactions among all of the heritable and non-heritable nodes.

My friend Dr. Wade (Indiana University, Bloomington) uses the metaphor of card games.  First, consider the child’s game of war in which each card has its own value.  In the game of war each player turns over a single card, and the highest card wins.  In this case we can, for example, say that being dealt an ace of spades is always good.  Contrast this with the game of poker.  In this case it is the value of the card in the context of the other cards in the hand that is important.  The ace of spades would indeed be a good card to get if the other cards in the hand were the ten through king of spades (making a royal flush); however, it might not be such a good card if the other cards were the 5 through 8 of diamonds (making nothing).  In both cases the card conveys the same information.  In the case of poker, but not war, the meaning of that information changes depending on the context (the genotype) of the other cards in the hand (modified from a forthcoming book by Michael Wade).

Genotypes are much more similar to a poker hand than to a war deck.  That is genes convey information, but the meaning of that information cannot be interpreted outside of the genotypic concept in which it is found.  However, the “five-card stud” version of poker I described above is also simplistic.  A better metaphor might be Texas holdem.  Texas holdem is a variant of poker in which (I only played this once or twice, so I am reading the rules) two cards are dealt face down to each players, and ultimately five cards are dealt face up, and shared by all players.  The goal is to make the best five-card hand out of the seven cards available to you.  In this metaphor the face up cards are the environment, and affect the value of everybody’s hand. The final hand (phenotype) is a mixture of the genotype (the down cards) and the environment (the up cards).  Importantly, the value of a down card (gene) is entirely dependent on the expression of the up cards (environment) and the other down cards (the genotype).

Regardless of how apparently obvious the above metaphor may be, it is an example of why it is important to make a distinction between the information that is assembled (e.g., genes and environment) and how that information is actually used in the patterning process that creates the phenotype.

The second reason for this partitioning, is that there is a tendency to distinguish between “genes” and “environment”.  One problem with this is that there are many things that can cause heritability, and the phenotype will be affected by all of them.  In addition to genes there can be inheritance of environmental factors.  To take a human example, an individuals income is strongly affected by their parents income.



Lest one think that this is a short term affect, the wonkblog (http://www.washingtonpost.com/blogs/wonkblog/wp/2012/10/18/how-your-last-name-will-doom-your-ancestors-centuries-from-now/) describes a study in which such wealth effects are shown to last for centuries.  In the figure below the Noble’s took over the Andersons (names from Swedish nobility) over 1000 years ago, an advantage that they still maintain.  Thus, to pretend that in humans inherited wealth and social standing are not important influences on the phenotype makes no sense.



Finally, it is worth noting that cultural inheritance is seen in animals other than humans as well (http://www.nytimes.com/2013/04/26/science/science-study-shows-monkeys-pick-up-social-cues.html?_r=1&)

The other aspect of this problem is that the “environment” enters into evolution in two manners.  First, aspects of the environment influence the phenotype, as I have been discussing in this post, and second, fitness is determined by the interaction of the phenotype with the environment.  It makes it much easier to distinguish these two uses of “environment” if I call the first non-heritable nodes and the second environment or environmental forces.

It is tempting to think of this as a very depressing view in which nothing can be known about anything – that is, it is a complex system interacting in a complex manner producing a complex and unpredictable outcome.  However, one of the interesting features of complex systems is that they often have simple behaviors.  Under most circumstances this underlying complexity will give rise to simple behaviors that can lull us into the belief that the phenotype can be described by simple genetic systems.

The phenotype and evolution (more on defining evolution)

In the previous post I made it clear that I was defining evolution in terms of changes in phenotype.  This is an important point, as many only consider changes in gene frequency to be evidence of evolution.  As an example, in a recent article in the scientist (http://www.the-scientist.com//?articles.view/articleNo/35317/title/Humans-Under-Pressure/) the author writes: “Although the team did not find any actual changes in gene frequencies—the gold standard for demonstrating evolution has taken place . . .”.  This would seem to imply that it is genetic change, rather than phenotypic change, that is the important sign that evolution has occurred.  There is a logical fallacy here: While a change in allele frequency necessarily means that evolution has occurred, the opposite is not the case.  Documentation that evolution has occurred does not necessarily mean that allele frequencies have changed.

To see this, consider the change in stature of the Japanese people since 1870 (figure modified from http://www.dh.aist.go.jp/en/research/centered/anthropometry/).  Since 1870 there has been a secular change in the stature of the Japanese from roughly 5 feet tall in 1880 to the modern height in 1980 of roughly 5 feet 5 inches  (that is averaging men and women).  It is safe to say that little if any of this change in stature is due to a change in allele frequency.  The rather obvious conclusion is that this change in stature is due to a change in diet from a largely starch (rice) based diet in the 1800’s to a modern diet rich in proteins and vegetables.  The best data on this are shown in the dietary changes since the end of world war 2 (from www8.cao.go.jp/syokuiku/data/eng_pamph/pdf/pamph3.pdf). 

Height of Japanese over time Japanese diet

The question is, is this evolution?  Recognize that the heights given in the graph are adult heights, thus, I can reasonably argue that these are adult phenotypes, that will be subject to at most minor changes within individuals.  In short these are not developmental changes.  Thus, we see the population getting larger due to the birth of individuals that will grow to be tall, and the death of older Japanese that were shorter.  This fulfills my definition of evolution, and as a consequence it seems to me we must call it evolution.  Note that if you do not accept my definition, it is also evolution under Futuyma’s definition.  That is it is a lasting change in mean phenotype that transcends the life of an individual.

The temptation is to say that this is “simply” a change in environment.  However, it is better considered a change in the cultural milieux.  Changes in diet are due to the changes in the behavior of individuals, which is learned, and culturally transmitted.  In short, it is heritable.  From this it should be clear that the evolution, that is changes in the distribution of the phenotype, need not be supported by genetic changes.  Anything that can contribute to secular changes in the phenotypic distribution can support evolutionary change.  (Note, I was going to say the resemblance between parent and offspring, but with culture such simple statements are fraught with peril –see Boyd and Richerson’s books http://www.amazon.com/Not-Genes-Alone-Transformed-Evolution/dp/0226712125/ref=sr_1_1?ie=UTF8&qid=1367256338&sr=8-1&keywords=boyd+Richerson)

Much more complicated, in my mind is whether secular changes in phenotype due to secular changes in the environment count for evolution.  I know of no clear examples of this, so a hypothetical example will have to suffice.  It is known that snakes grow larger in warm environments, thus, it is reasonable to imagine that the average size of garter snakes will increase as a result of global warming.  This change could involve simply changes in the environment (mean temperature), and nevertheless would satisfy both my definition, and Futuyma’s definition of evolution.  It seems to me that if it satisfies the definition of evolution we must consider it to be evolution.  I will say I am not fully happy with this, but well, it is what it is.

Addendum:  As evolutionary biologists we often find ourselves fighting the fight of evolution vs creation.  I think that this has the effect of Balkanizing our thinking.  This is particularly true of our thinking of what is and isn’t evolution.  I am purposely writing this blog to be read by evolutionary biology professionals, and as such trying to push my thinking to the logical ends, whether or not it makes it difficult to hash over stale arguments about whether or not Jesus petted dinosaurs (by the way:  Assuming he kept chickens, and given we know that birds are Saurischians, yes, he probably did pet dinosaurs).

Defining Evolution

One of the unusual things about books about evolution is how frequently they fail to actually define evolution.  Typically such books will have several pages describing evolution, so that by the time you have finished reading the section you have a pretty good idea of what they are talking about.  But generally speaking if after reading one of these sections you were asked to complete the sentence “Evolution is defined as . . .” you would be unable to do it.  The hesitance to define evolution is quite understandable.  Indeed, the word “evolution” does not appear in Darwin’s Origin of Species.  Instead, he uses the term “descent with modification”, a phrase that remains as a good definition of evolution.  Darwin’s idea of descent with modification emphasizes both the important aspect of change (modification), but also the idea of transmission between generations.  That is, change within an individual is not enough, rather it is change among individuals, and more specifically change that persists across generations.

Darwin’s definition raises the important point that any definition of evolution has implicit in it the concept of change, but the context in which that change is described is an essential element of the definition.  I am told that this is the reason that Darwin avoided the term evolution.  At the time he was writing his book evolution typically was used to describe the change in an human as they matured, or otherwise went through intellectual or spiritual changes.  If indeed, in Darwin’s time “evolution” referred to a theory that an embryo was a growing or unfolding from a preexisting form with rudimentary parts of the future organism.  At the time evolution would have also been used to refer to the orderly unfolding of events, such as the intellectual or physical developmental changes within an individual over time.   Because of this, he was absolutely correct to avoid using that term.

In common language we are quite comfortable with the thought that a persons opinions can evolve, however, this is not compatible with the generally held concepts of organic evolution.   In organic evolution it is generally held that it is a set of objects that evolves or changes, and that the set changes through the addition, subtraction, or replacement of the objects that make up the set.  In typical thinking about evolution the set is a population and the objects are organisms, however, this need not be the case.  There is no reason why the objects cannot be cells, and the set an individual organism.  Conversely, the objects could be populations and the set could be a group of populations, or as it is typically called, a metapopulation.  Indeed, evolutionary change can be described for any entity that can be described as an object, and is contained within any group of those objects that can be described as a set.  The important point is that in common thinking change in the objects is not considered to be evolution.  This is not to deny that objects change.  An organism is a typical object within a group that is a population.  Organisms are born, grow, mature, and eventually senesce and die.  However, these changes are not considered evolution, rather we give them a different name.  For organisms we would call these changes “development”.  If the object were a population within a metapopulation then we might consider the changes that take place within the population to be development, ecological changes or demographic changes.  However, again, we would not consider the changes within the object to be evolution.  This is a subject I will discuss in a later post.

Darwin deals with the distinction between change within objects (development, ecology, demography) versus the change between objects (evolution) by emphasizing the concept of descent.  Descent with modification alludes to the fact that organic evolution involves objects that reproduce in some form, and generally have a finite life span.  Thus, new objects enter the group through some form of reproduction, and old objects leave the group through some form of mortality.  Evolution occurs because the new objects entering the group are in some way different from the old objects that are leaving the group.

Over the years many definitions of evolution have been presented.  One of the more famous suggested definitions is that evolution may be defined to be a change in gene frequency.  I cannot find the original source for this, but it clearly dates from the 1930s and 40s, that is the time of the “new synthesis”.  The new synthesis was a theoretical consilience of the (at that time new) field of genetics with the changes in appearance, behavior, and physiology of organisms that scientists had been studying for years.  One of the main fields to come out of this synthesis was the field of population genetics, which is the study of gene frequencies in populations.  Seen in this light it is quite clear that by this definition that genes are considered to be the determinants of phenotype, thus the “modification” in descent with modification are changes in the frequencies of the underlying genes.  The implication being that phenotypic changes could ultimately be traced to genetic changes, thus it is reasonable to define evolution strictly in these terms.

There are, however, two issues with this gene-centric definition that make it inadequate.  The first is the obvious issue that this definition eliminates any changes that do not have a genetic basis.  The most obvious of these is “cultural evolution” that is changes that are learned from other members of the population.  Such changes need not, and in many cases don’t, have a genetic basis, and yet they lead to changes in a population that can persist for many generations.  Another interesting example is changes in a population due to the prions.  Prions are alternate conformation of otherwise normal proteins.  In their normal conformation these proteins perform a normal cell function, in their alternate conformation they are the causes of diseases, the most important example of which is spongiform encephalopathy, or so called mad cow disease.  A protein takes on this alternate disease conformation only in the presence of other proteins that are also in that disease conformation.  Without going into detail about these diseases, the point is that prions do cause changes in populations, and these changes are based on conformational changes of proteins rather than changes in the frequency of alleles.

The second issue is subtler.  In standard genetics text books there are generally three domains that are identified.  These are the phenotype, the genotype and the genes.  The phenotype is the appearance of an organism broadly defined.  That is the phenotype is all aspects of the morphology, behavior and physiology of an organism.  The genotype is the genetic makeup of an individual, and finally the genes are the genes that are present in the population.  For the present discussion the important aspect of the phenotype is all aspects of the morphology behavior and physiology of an organism. Thus the genes that make up an organism belong to the phenotype to the extent that they are part of the “morphology” of that organism, however they are part of the genotype to the extent that they cause other aspects of the phenotype. Thus, the sequence of nucleotides in the DNA is part of the phenotype, however the effect of a gene on other aspects of the phenotype (say hoof shape) is not part of the phenotype.  The important aspect of the genotype is that it is the specific assemblage of genes in an individual that interact with the environment to produce the phenotype.  Thus, we need to recognize that it is the genotype in its interaction with the environment and things like culture that determines the phenotype.  Finally, the genes are important because it is these that are passed from parent to offspring.  Thus, the resemblance between parents and offspring is to a major extent due to the transmission of genes.  The genes, as such, to not determine what a trait will be.  It is only after they are assembled into a genotype that a gene affects the phenotype.  Ignoring cultural transmission, offspring resemble their parents only to the extent that the genes transmitted by the parent influence the genotype to affect the phenotype in a predictable manner.  Thus, the genes themselves have no influence on the phenotype, rather they are assembled into the genotype, and it is the genotype that influences the phenotype.  The point is that the gene domain is important in the transmission of traits between generations, not about the expression of the phenotype.

When Darwin spoke of descent with modification he was referring to changes in phenotype, and we should not lose sight of the idea that evolution is about changes in phenotype.   Defining evolution as change in gene frequency misses the fundamental point that evolution should be defined in terms of phenotype, the domain that is changing, rather than in terms of genes, the domain that is mediating that change.

A second definition I will consider is one provided by Futuyma in his book Evolutionary Biology.  In this book he defines evolution as “lasting change in the mean phenotype of a population that transcends the life of an individual”.  This definition is actually quite similar to Darwin’s in that it focuses on changes in phenotype, and specifically excludes changes within an object.  The first clause of this definition, lasting change in the mean phenotype of a population, emphasizes that indeed evolution is about changes in phenotype, but further demands that those changes be of a more or less permanent nature.  That is, this definition specifically excludes temporary changes due to short-term environmental effects.  As an absurd example, the fact that leaves change from dry to wet during a rainstorm is not an example of evolution.  The second clause, that change must transcend the life of an individual, specifically excludes developmental change.  In North America many insects over-winter as eggs.  In the spring the eggs will hatch, the insects will emerge, develop and mature into adults.  Over the course of a season the population will change from a population of eggs to a population of mature flying adults.  The second clause tells us that this is not evolution since it is a result of developmental changes taking place within an individual.  Thus, whereas Darwin’s definition emphasizes that evolution is change among units by emphasizing the reproduction and replacement aspect of living systems, Futuyma’s definition emphasizes that evolution is among objects by specifically excluding change within objects.

It is also interesting to note that Futuyma specifically identifies the objects to be individuals and the groups to be populations of individuals.  By individual it seems that he is referring to organisms.  However, as I suggested above, there is no reason that the objects cannot be cells and the groups organisms, nor is there any reason that the objects cannot be populations and the groups metapopulations.  There are two ways that this problem can be handled.  One is to change Futuyma’s definition to use a more generic term than individual.  The other is to decide that individual does not necessarily refer to organism.  The question becomes whether and under what circumstances it is reasonable to refer to a cell within an organism to be an “individual”, and conversely when a population within a metapopulation can be considered an “individual”.  I will address this in a later post.

Finally, a third definition of evolution that is frequently used is heritable change in the mean phenotype of a population.  This definition is similar to Futuyma’s definition in that it does emphasize the change in the mean phenotype of the population, and the term “heritable” is intended to separate within object developmental change from changes taking due to changing the objects in the set.  However, this definition falls short in the question of what does heritable mean.  Heritable in its most basic form means that the offspring resemble the parents.  The choice of this term over the gene-centric concept of change in gene frequency is good, in that the specific mechanism causing the resemblance is not specified.  However, this definition emphasizes that only changes that can be passed from parent to offspring count as evolution.  The problem comes when a trait is not completely heritable.  Consider the situation when the resemblance between parent and offspring is only 50%.  In that case we would certainly believe that change in the trait is evolution.  On the other hand if the resemblance between parent and offspring is 0, then the trait is not heritable, and we would not consider change in that trait to be evolution.  The question now becomes how low does the resemblance between parent and offspring have to become before we don’t consider it evolution?  That is, if the resemblance between parent and offspring is only 0.01% (which in any reasonable experiment would not be detectable) is that evolution?  Unfortunately, this definition of evolution puts an arbitrary break point in what is actually a continuous range of possibilities from traits not being heritable to traits being highly heritable.

This leads us to what is an appropriate definition for evolution.  I would argue that both Darwin’s definition of descent with modification, Futuyma’s definition of lasting change in the mean phenotype of a population that transcends the life of an individual, and heritable change in the mean phenotype of a population are adequate.  However, we need to recognize that evolution need not change just the mean of a population.  One of the forces of evolution is selection, and one form of selection is stabilizing selection.  Stabilizing selection does not change the mean of a population, rather it reduces the variance in a population.  This occurs because extreme individuals do not survive and reproduce as well as individuals near the optimum, which is usually very close to the population mean.  Thus, we should probably consider any changes in the distribution of a population, not just changes in the mean, to be evolution.  I would go on to argue that neither definition explicitly states that evolution takes place by replacements.   Thus, I would first argue that Futuyma’s concept of change in the mean phenotype needs to be expanded to include any changes in the distribution of phenotypes, including both changes in the mean and changes in variance.  Second, we need to explicitly recognize the idea that our concept of evolution is change due to the turnover of objects, not due to the change in those objects themselves.  Thus, I suggest that a first pass of a definition of evolution is change in the distribution of phenotypes in a set due to the addition, loss or replacement of objects.  If we decide that “sets” always refers to some form of population, and that objects can always be called “individuals” then we get a less clumsy definition that uses a language that is consistent with other definitions.  Replacing set with population and object with individual we can define evolution as change in the distribution of phenotypes in a population due to the gain, loss, or replacement of individuals.  This definition conveys nearly the same information as Darwin’s and Futuyma’s, however it makes explicit that evolution is due to a change in the makeup of a population through the gain and loss of individuals.


Thus, we can finish the sentence posed at the beginning of this post:

Evolution is defined as the change in the distribution of phenotypes in a population due to the gain, loss, or replacement of individuals.

Introduction to this blog

I started studying evolution when I entered graduate school over 35 years ago.  Since that time I have found that my views on evolution are often at odds with those of other often very outspoken evolutionists.  I also find that my views tend to be in line with the work-a-day evolutionists.  In other words, the often glib evolutionists that the lay public is familiar are often telling a simplified story that has little to do with the reality of those who actually look at data and develop useful models.

Because of this I am frequently informed that I need to write a book or a review article of some such giving an enlightened view of evolution that is more in line with that of a large but under appreciated group of scientists.  I have avoided doing that out of abject fear.  Many of the popular evolutionary biologists are famous for making ad-hominem attacks, and for being quite vicious in their attacks.  As a junior scientist not at one of the major research institutions self preservation made me shy about expressing my views in public.

That, has changed as I move into the later half of my career and have no intention of moving from my current position.  My goal in this blog is to start laying out some of my ideas on how evolution works, with the eye towards perhaps turning some of these essays into a book at some point in the future.

Skip to toolbar