Lewontin has a great grasp of the machine metaphor in biology, going so far as to characterize the primary metaphor of molecular biology as such, then further characterizing developmental genetics initially with cybernetic control and currently with electro-mechanical systems, in successive attempts to integrate the notions of signaling processes and regulatory feedback, which have replaced the original clock with gears metaphor.

“Remember that no matter how selfish, how cruel, how unfeeling you have been today, every time you take a breath, you make a flower happy.” - Cute quote from Mort Sahl

Misc notes, the last two being the most interesting

• He posits that in the case of ecology and evolution, the metaphor is that the environment posed “problems” while the organism proposed “solutions”.

  The current dilemma for these metaphors is the notion of multiple sources of cause and effect, citing the co-dependencies of organism and environment in their evolution.

• He provides a very nice overview of what he sees as the three major stances towards research in biology, and relates them to the current practice of reductionism.

  1. Gaia theory: If you devise a graph relating all living things, you get one connected component. In other words, everything is connected to everything else, and the appropriate form of “life”, as a singular noun, is a massive cybernetic control system, with no components that can be analyzed independently or out of context of the whole. In effect the only way to study this system is to consider the whole. No reduction is possible.

  2. Contextual reductionist theory: Living things are connected to each other in unique and heterogeneous ways, and the best way to understand systems is to identify the various boundaries of components. That is to say, the graph of life has many connected components, and biology is best studied by analyzing each component as the whole. This approach is much simpler than considering all life as the whole. He claims this is more or less the goal of the current reductionist scheme, with the exception that in biology there is no clear picture of the whole, and so one must start organizing a list of parts. Once this is achieved the goal is to piece together only the parts that interact meaningfully as a connected component. Lewontin believes this is the bulk of what needs to be done for future reasearch in biology. The devils are in the details, and there are lots of devils to hunt down.

  3. Generalisms: the connected components can be identified by local edge densities, i.e. various properties of the graph itself. From this perspective you get things like applications of graph theory to biology, scale free networks, and Kauffman’s NK models and boolean networks.

• Interesting overview of three attempts to identify and characterize the laws or generalizations of biology

  • René Thom’s Catastrophe theory
  • Chaos theory, deriving from attempts to predict storm patterns and various chaotic regime dynamical systems
  • Complexity theory, with the exemplar of Stuart Kauffman’s Origins of Order

  Interestingly, Lewontin seems to generally dismiss these attempts at generalizing biology, on the grounds that living systems are too heterogeneous to be so characterized. At the same time he at least admits that Complexity theory has not been showed to be a fruitless endeavor yet.

  In addition there is no mention of cellular automata or Wolfram’s ANKOS. I am curious whether this reference is simply agglomerated into Complexity theory, or not sufficiently biologically focused to merit attention.

• Lewontin places his eggs in the basket of shape, form, and structure, claiming that future progress in biology is precluded without an understanding of how the spatial context of a cell, and the organelles and genes and proteins within it, is used to connect gene expression to macromolecular state.

  • “The problems of cell differentiation, division, and movement cannot be solved without information about the spatial distribution and organization of molecules within cells and about how the state of a cell is influenced by neighboring cells and the surrounding environment.”

• Quantitative trait analysis requires understanding the internal developmental and external environmental factors that affect the generation of phenotype, given a genotype (understand the reaction norm and developmental noise).

• Emphasis on the importance of small molecular effects combining synergistically to generate macromolecular structural change (i.e. evolution of form and function).

  • Leading example is the extreme purifying selection on genes. You expect 3/4 of coding mutation to be nonsynonymous, but find only about 0-10% change.

• Lewontin appears to support the neutral theory as a move influential evolutionary force than positive selection. Also he claims that the majority of between species adaptation results from such neutral changes. Thus he claims that the study of variation within species is a much more reliable guide to the organismal level function of molecules.

• Two main features that distinguish living things from other physical systems

  1. Observed variations in the behavior and morphology of individual organisms in nature usually have no consistent effect on function when averaged across contexts, in contrast to variations in molecular structure. His argument is that according to the theory of natural selection, any genetic variation that has a consistent effect will have a strong selective effect, and so this variation will be quickly consumed in the process of evolution and the relevent alleles driven to fixation. Any remaining variation in the trait will be caused by developmental or environmental noise, or is effectively selectively neutral.

  2. Living systems are open, engaging in exchange with the external environment. Organisms and environments are both causes and effects in a co-evolutionary process. The one general feature of this process is that it is topologically continuous, meaning the evolution of either can be tracked by following the small scale changes in one or the other (environmental change covaries with organismal change). Organismal selection is studied in three ways 1. Measure fecundity and viability in a particular environment 2. Observe fitness changes of given genotypes as environment changes 3. Follow gene frequency changes in a population over time, or differences in changes between populations as related to external environmental conditions.

     More should be done to track the converse, how environmental selection changes over time. If you can’t measure the variables directly, use the property of continuity and use the frequency changes in the organism to indicate an environment altered by the organism.