Thursday, April 10, 2014

More than just a metaphor, Wright’s Adaptive Landscape provides inspiration

[Originally appeared on Nothing in Biology Makes Sense]

Review of The Adaptive Landscape in Evolutionary Biology edited by Erik Svensson and Ryan Calsbeek

Have you ever wished you could go back in time to be present at a particular historical event? The 1932 International Congress of Genetics sounds perfect, right? There R. A. Fisher, J. B. S. Haldane, and Sewall Wright all presented papers of their recent research. If you’re a student of population genetics, you probably recognize these names as some of the founders of the field. At this meeting, Wright was asked to condense some of his more technical mathematical framework into a form that was more widely accessible to the audience of biologists. The result was his conceptualization of the Adaptive Landscape where an analogy is made between the fitness of an individual or population and the varied topographic landscape (pictured on the cover of the book). Wright used this metaphor to describe aspects evolutionary dynamics of populations.

The editors of a recent book, The Adaptive Landscape in Evolutionary Biology, gathered together contributions from evolutionary biologists, ecologists, and philosophers to demonstrate the impact that the Adaptive Landscape has had on the field of biology. This book embraces an 80 year old metaphor created by one of the founders of the modern synthesis to explore the breadth and depth of research generated in evolutionary biology. Unlike a recent book addressing aspects of the modern synthesis, Evolution: The Extendend Synthesis (Pigliucci and Müller, 2010) which called for a revolution, Svensson and Calsbeek have assembled authors that explore the innovations and contributions that build upon the fundamental ideas of population genetics and seek to grow the field. Early in this book, Pigliucci asks about the utility of the Adaptive Landscape metaphors, even titling his chapter with the question, “what are they good for?” I think the rest of the book provides a more than sufficient answer to his question.

The book in five parts: The book is broken down into five parts following a logical progression of information. The first part, Historical background, gives you a good sense of where the Adaptive Landscape metaphor came from as well as quickly showing you how future authors adapted and modified the heuristic for other purposes. The second part, Controversies, includes contributions from three theoreticians familiar with the mathematics and assumptions behind the Adaptive Landscape metaphor. The third part, Applications, is the largest portion comprising about one third of the content and covers various empirical examples with detailed analysis sections helpful to those interested in the specifics. The fourth and fifth parts, Speciation and Development respectively, cover some disparate examples but force the reader ton consider a macroevolutionary perspective.

Over the course of this semester, my colleagues and I read and discussed each of the chapters. Our group consisted of a diversity of backgrounds spanning evolution, ecology, and behavior. We included a range of experience from first and second year graduate students to postdocs as well as junior and senior faculty. While we read the book in the sequential order it was published in, each part could certainly be pulled out and read as a separately. Some of the chapters make cross references to each other but not enough that reading them independently would be impossible. While some chapters certainly provided more challenges to some, this forced our discussion to flesh out explanations that the text just didn’t have time to go into. I think that the first two parts might make particularly good set of readings to supplement an upper level Evolution or Population Genetics course.

My recommendation

This is a book ideally suited for a graduate level seminar in any Biology department. While it may be good to read a few chapters on your own, the book benefits from an active discussion of the content. We took 14 weeks to go through the entire book, reading one chapter most weeks occasionally two (or three). This pace allowed everyone to casually complete the readings. Going slowly through the book also allowed us to digest the material and make connections among the chapters without getting too overwhelmed with new information each week. You can follow this link to see the schedule we followed.

If you are not convinced about the utility of this book yet, below I highlight some of the excellent contributions contained within the different parts.

 Wright’s original drawing of the Adaptive Landscape

Historical Background: The first and second chapters are nicely dovetailed with one another and provide a more than complete history of the origin story of the Adaptive Landscape metaphor. This section includes background of potential parallels to earlier ideas from a French engineer, but I’ll leave it to you to discover those mysteries in reading the book. Skipper and Dietrich do spend some time on the differences in the interpretation of the Adaptive Landscape for those interested in the philosophy of biology. This section brought up some potential pitfalls of over interpreting the Adaptive Landscape. As a biologist with some confidence in mathematics, I appreciate a good figure, but one must remember it has limitations. Don’t forget the mathematical model underneath that beautiful visualization.

Controversies: Three prominent theoreticians in the field of population genetics provide the contributions for this part. Steven Frank dives into the well know argument between R. A. Fisher and S. Wright. An interesting bit of historical context in the battle between Fisher and Wright is that it may have been more one sided than you might think. It seems that that Wright wrote more about Fisher’s theory than Fisher often framing the argument to his advantage. Frank provides a well-reasoned description of Fisher’s fundamental theorem where he contrasts the change in population fitness due to natural selection and that due to change in the environment.
As a side note, Frank has been writing a series of publications on natural selection which may be of general interest to students of population genetics and can be found here.

Wright’s metaphor of the Adaptive Landscape was originally intended to provide a means to understand his Shifting Balance Theory. Wade outlines the theory behind the three phases of the Shifting Balance Theory along with providing copious empirical examples from his own work along with others. (Full disclosure here, Michael Wade, chapter 5 author, is my current postdoctoral advisor but holds no editorial review on this post.)
 Contained within this chapter is an excellent analogy of the difference between additive and epistatic effects on fitness. Think of the two classic card games: War and Poker. One can think of additive effects as though one were playing War where the outcome of a single round is dependent the face value of a particular card regardless of the context. On the other hand, in Poker, the value of a particular card depends on the context of the other cards in the hand. Sometimes an ace is very strong, for instance when paired with three others. However, when you are trying to complete your two-X-four-five-six straight, an ace has little value. Just like in poker, with epistatic fitness effects, the context of an allele (or card) matters.

Goodnight wraps up the controversies portion of the book by diving into the underlying assumptions of the mathematical model for Wright’s shifting Balance Theory. He does an excellent job of addressing some of the critiques that authors from earlier chapters posed regarding evolution across an Adaptive Landscape. Goodnight also spends some time with Gavrilets’s alternative metaphor, the holey landscape. While many authors in this book include ideas and contrasts from the holey landscape, I wish that Gavrilets had written a chapter for this volume even though he has already written extensively on this subject (Gavrilets, 2004).

Applications: As one might expect, the bulk (over one third of the entire book) covers applications of the Adaptive Landscape. I think you could argue that Parts IV and V could be covered within this section as they include many applications. This part of the books includes a very excellent chapter (ch. 9) on analyzing and comparing individual fitness surfaces based on empirical data. While note quite a data analysis walk through, with a little extra help this chapter certainly provides you with enough of an outline to help design your own empirical investigations. The boxes within the text are very clear examples of using the now classic regression approach (Lande and Arnold, 1983). Several of the chapters within this part address the important idea that a fitness landscape may not be static but may vary across space and even time. Outside of this book, the idea that the form and shape of the fitness function may vary is also discussed at length in studies of coevolution, specifically selection mosaics (Gomulkiewicz et al., 2007,Thompson, 1999, 2005).

Speciation: Part IV of the book addresses the immense and challenging world of macroevolution. These chapters describe some of the potential challenges of incorporating evidence spanning the history of life on earth. At first, their message may seem overly cynical because of the large uncertainty (at multiple scales) associated with the fossil record, however, they do provide some thought provoking conclusions about the state of current research. This is one section where the evolution or change of the adaptive landscape is particularly emphasized as researchers are dealing with extremely long periods of time. It will be very interesting to see where the research is heading as computational power increases and model reconstructions of the past become more and more accurate. The brief introduction to the Adaptive Dynamics modeling framework (chapter 14) does not seem all that well integrated into the overall theme of the book.

Development: This part of the book brings together a heterogeneous set of topics. I found the empirical examples of butterfly mimicry intriguing, but felt as though the new metaphor that Wagner introduces in chapter 17 lacked the helpful intuition that the original Adaptive Landscape provides. Sean Rice provides a rather thorough dive into the complex world of mapping abstract underlying factors to phenotypes as well as phenotypes to fitness in the context of development. While Rice’s “Formal analysis” section of his chapter may be beyond the average reader, he provides an excellent intuition in the “informal theory” section as well as interspersed throughout the chapter.

Sewall Wright (note the guinea pig)

CONCLUSION: I would strongly encourage students of population genetics to pick up this excellent volume and spend some contemplative weeks reading through the chapters. Better yet, grab a group of your department colleagues and argue about the 80 years of interpretation of the Adaptive Landscape. I personally cannot wait to see where this excellent metaphor leads us.

  • Gavrilets S (2004) Fitness Landscapes and the Origin of Species (Mpb-41). Princeton University Press.
  • Gomulkiewicz R, Drown DM, Dybdahl MF, Godsoe W, Nuismer SL, et al. (2007) Dos and Don'ts of Testing the Geographic Mosaic Theory of Coevolution. Heredity 98: 249-258. DOI: 10.1038/sj.hdy.6800949
  • Lande R, Arnold SJ (1983) The Measurement of Selection on Correlated Characters. Evolution 37: 1210-1226. DOI: 10.2307/2408842
  • Pigliucci M, Müller GB (2010) Evolution: The Extended Synthesis. Mit Press.
  • Svensson E, Calsbeek R (2012) The Adaptive Landscape in Evolutionary Biology. OUP Oxford.
  • Thompson JN (1999) Specific Hypotheses on the Geographic Mosaic of Coevolution. American Naturalist 153: S1-S14.
  • Thompson JN (2005) The Geographic Mosaic of Coevolution. Chicago: University of Chicago Press.
  • Wright S (1932) The Roles of Mutation, Inbreeding, Corssbreeding and Selection in Evolution. Proceedings of the Sixth Annual Congress of Genetics 1: 356-366.

No comments:

Post a Comment