The Arrival of the Fittest: a book review

Andreas Wagner’s book,The Arrival of the Fittest has been mentioned many times (just try a site search as I’ve just done) since it was published. Petrushka pointed it out in a comment

For anyone interested in whether RMNS can create stuff, I recommend a relatively new book, Arrival of the Fittest. I just bought the Kindle version an haven’t finished, but it has a lot to say about how goldilocks mutations occur.

Much later Mung writes:

Reminds me of petrushka, who is always plugging Andreas Wagner’s Arrival of the Fittest, but will never post an OP on it for discussion.

So I’ve taken the hint and bought the book at last. I can see why people have recommended it.

Wagner writes clearly and fluidly. The first chapter is an outline of the history of evolutionary biology, from Darwin to today, and the related fields of genetics, biochemistry and molecular biology. He considers the modern synthesis and how, while hugely successful in dealing mathematically with the genotype and selection, fails to get to grips with the phenotype and sources of innovation. Hopes were raised with the new discipline of evolutionary developmental biology (evo-devo) but Wagner observes:

Where the modern synthesis has a theory without phenotypes, the embryologists have phenotypes without a theory.

Wagner suggests biochemists have begun to bridge the gap by showing that proteins build the phenotype. Wagner mentions the fact that whilst we now have the technology to sequence genomes and construct 3D models of proteins in complete atomic detail, we as yet cannot predict the folding and functionality of novel protein sequences. Later in the chapter, he remarks that unlike, say, advances in physics, advances in evolutionary biology have left the core concepts intact. (Professor Moran may disagree!)

Chapter 2, entitled The Origin of Information, is a review of “origin-of-life” hypotheses. Wagner favours RNA world, hydrothermal vents and the citric acid cycle as primordial. He points out that the citric acid cycle produces a lipid precursor; lipids, he points out, having the emergent property of self-organising into membranes and vesicles in an aqueous medium. Continuing the chemical pathway theme, Wagner moves on to the universal energy molecule, adenosine triphosphate (ATP). He ends the chapter by listing his list of essential ingredients for life to emerge from non-life. First, some chemistry needs to happen prior to becoming life – involving simple catalysts such as metal complexes and then we need innovative reactions, innovative biological catalysts to speed those reactions and regulatory elements to coordinates the reactions.

Chapter 3, The Universal Library, discusses cell biology and the fact that at least 5,000 different chemical reactions occur in at least one organism, somewhere. No single organism performs all 5,000 but some bacteria perform many. E. coli manages more than a thousand. Wagner goes on to explain the idea of the metabolic genotype, a comparison based on which metabolic pathways an organism is able to perform and the “geniuses” of innovative metabolic pathways are prokaryotes. Wagner attributes this to horizontal gene transfer being rife across bacterial species and their very rapid generation time (twenty minutes for E. coli).Wagner points out the variability in stains of E. coli which can be 25% of their genome. Wagner goes on to describe his research using computer analysis that shows the staggering variation in closely related bacteria.Wagner finds there are many solutions to a metabolic pathway, rebutting the “needle in a haystack” canard.

Chapter 4 is entitled Shapely Beauties and is an in-depth look at cell biochemistry for the interested layman. He touches again on the “needle in a haystack” argument when discussing finding functional proteins in among the theoretical number of all possible proteins and describes Keefe and Szostak’s experiment to find functional proteins. He mentions that insects and plants both have oxygen-binding globins that have similar shapes and folds, do similar jobs, yet differ in 90% of their amino acid sequences. There’s also a digression into RNA world. He describes further work at his lab developing the concept of genotype networks.

Chapter 5, Command and Control,is about gene regulation and the field of systems biology which marries experimental data with mathematics and computing. Wagner talks about his collaboration with Olivier Martin, a statistical physicist. Again, he demonstrates the robustness of regulatory processes and the huge amount of viable variation.

Chapter Six, The Hidden Architecture, develops the idea of <i>robustness</i>, resilience in the face of change. For example, the protein lysozyme, contained in tears and saliva, is a bactericide. Scientists have engineered some 2,000 variants involving one altered amino acid and 80% of those variants still kill bacteria. Another example Wagner uses are crystallins, that form transparent lenses for eyes. Their refractive index is ideal for the job. Yet crystallins also perform metabolic functions. Wagner mentions the selectionist neutralist controversy and disputes Kimura’s suggestion that most genetic variation is neutral but agrees that some neutral variation is necessary to his idea of genetic networks.

The final chapter, From Nature to Technology, is on a theme of “trial and error” and the phenomenon of exaptation, where some organ or system becomes redundant for it’s original purpose and gets reworked for a new one. A digression into Boolean algebra, truth tables and computing was helpful to me but I probably need to read it again.

Then we come to the epilogue Plato’s Cave. Thankfully, no mention of Kairosfocus! It’s Wagner’s advocacy for mathematical modelling and computing as a tool that should bring powerful insights into biology.

Summing up, a good read, wide-ranging and informative for the lay reader. I recommend it.

111 thoughts on “The Arrival of the Fittest: a book review

  1. Allan Miller: It’s there in every argument. Evolution can’t do X, therefore …

    I think you left out the heart of the ID argument.

    Evolution can’t do X because X is really complicated, and we can actually put a number on the probability of X occurring, assuming evolution can’t do it.

  2. Interesting that Mung seems to have lost interest in this thread just after being asked to summarize some of the crucial ideas in the book.

    What’s up with that, Mung? Didn’t you understand the book?

  3. petrushka,

    Evolution can’t do X because X is really complicated, and we can actually put a number on the probability of X occurring, assuming evolution can’t do it.

    I occasionally had fun (we live in a rural area; we have to make our own fun) ragging Joe G about that. Evolution isn’t testable; ID is. Step 1 in ‘testing ID’: test evolution.

  4. keiths: Interesting that Mung seems to have lost interest in this thread just after being asked to summarize some of the crucial ideas in the book.

    My interest is still where it always was. I am interested in seeing someone present an argument from the book that poses a problem for ID.

    A hyperastronomical search space isn’t one of them.

  5. Mung,

    A hyperastronomical search space isn’t one of them.

    A hyperconnected one is – at least to the version that regards substantial movement through the space as forbidden – the ‘isolated islands of function’ brigade.

  6. Allan Miller:
    petrushka,

    I occasionally had fun (we live in a rural area; we have to make our own fun) ragging Joe G about that. Evolution isn’t testable; ID is. Step 1 in ‘testing ID’: test evolution.

    Joe’s reply is don’t blame ID for evolution’s failure to be testable

  7. newton,

    Joe’s reply is don’t blame ID for evolution’s failure to be testable

    Yeah, I’ve seen ‘Joe’s reply’ a thousand times. It’s like saying don’t blame me that a particular variable is always zero, but then going ahead anyway and using it as a multiplier in some fancy-Dan equation.

    IOW, if something is truly not testable, you can’t claim that something that crucially depends on it is [testable].

  8. Allan Miller: Yeah, I’ve seen ‘Joe’s reply’ a thousand times. It’s like saying don’t blame me that a particular variable is always zero, but then going ahead anyway and using it as a multiplier in some fancy-Dan equation.

    IOW, if something is truly not testable, you can’t claim that something that crucially depends on it is [testable].

    No argument here.

  9. Allan Miller: A hyperconnected one is – at least to the version that regards substantial movement through the space as forbidden – the ‘isolated islands of function’ brigade.

    A hyperconnected search space would be just as miraculous.

  10. From a week ago:

    Mung,

    Perhaps you could start your summary by

    1) telling us what “libraries”, “networks”, and “readers” are; and

    2) explaining how the networks factor in to the readers’ explorations of the libraries.

    I need to know that you understand these concepts before I can explain why Wagner’s book is such bad news for ID.

  11. Mung still seems to be spooked by the thought of summarizing.

    Meanwhile, I’ve posted some good ID-relevant quotes from the book on swamidass’s new thread here and here.

Leave a Reply