November 23, 2014

Ratchets, Constructions, Games, and Borg in the Reading Queue

Here are a few new (or new to me) papers that are evolution-related from my reading queue. There is a loose theme to these papers (indicated in the title of this post). I will give you my impressions and insights as the post proceeds.

Varieties of uni- and multicellular relationships amongst different species of green algae. COURTESY: Figure 1 in [1].

1. Libby, E. and Ratcliff, W.C.   Ratcheting the Evolution of Multicellularity. Science, 346, 426-427 (2014).

This short paper in a recent issue of Science deals with the transition to multicellularity and associated "ratcheting" mechanisms in what is a complexity theory take on evolutionary transitions and ratchets. According to Libby and Ratcliff's model, the transition to multicellularity involved a transfer of fitness costs from individual cells to groups of cells. This could also be seen as an overall change in the level of selection from individual cells to cell populations [2]. In this case, a so-called ratcheting mechanism is also proposed that provide a mechanism for how such transitions occur. Group living allows for certain group traits to emerge and limits reversion to the single-celled state, a so-called "de-Darwinization" of individual-level cell behaviors [3].

While individual cells transition from being autonomous to being mutually reliant, this occurs only in the context of its fitness effects. For the complexity ratchet to work, there must be opposite effects on fitness. Group living in the form of a colony formation provides a fitness benefit that is not at all present with individualistic cells. Social arrangements such as division of labor can encourage these fitness effects. The authors point to apoptosis as a trait which, while having a high fitness cost to the individual, can be beneficial to the group. Namely, relaxed group selection on apoptosis allows for the growth and nutrient constrains of a population to be circumvented. There are other traits for which individual and group selection differ -- changes in these selective pressures (much like what one would see in a shift to a new environmental niche) are what drive the evolutionary transition.

Do cultural practices (such as dietary innovations) have an influence on human evolution? COURTESY: Frank Stockton, Smithsonian Mag.

2. Laland, K.N., Odling-Smee, J., and Myles, S.   How culture shaped the human genome: bringing genetics and the human sciences together. Nature Review Genetics, 11, 137-148 (2010).

Transitions to group living also involve new sources of fitness costs, distinct from those that exist at the individual level. In this lengthy review article (from 2010), Laland, Odling-Smee, and Myles argue that culture can modify fitness costs for a given trait. The article authors are also advocates of niche construction theory and a post-synthesis evolutionary theory [4], which is clearly seen in their treatment of how culture interfaces with evolution. The mediating effects of culture on population genomics and evolutionary dynamics can be seen in gene-culture coevolution, but also suggests that there is another dimension to group selection in animal species that possess culture. These authors might also argue that cultural selection pressure is a key factor in human uniqueness, something we will come back to in the next section.

While the examples given in the article are muddied with more standard environmental selection pressures, the argument for cultural selection is that some environmental pressures are specifically related to cultural construction [5]. For example, natural selection due to dietary practices are reinforced by human modification of the environment (e.g. harvesting animal milk for general consumption). While genes and culture are viewed as interacting forms of inheritance, it is not so clear as to how they can be disentangled. In cases where cultural boundaries shape population genetics, the answer is relatively easy. However, in cases where cultural dynamics influence the frequency of host-pathogen interactions or heat shock genes, removing the signal from the noise is less clear.

Chimps playing the Prisoner's Dilemma. COURTESY: Science Magazine.

3. Martin, C.F., Bhui, R., Bossaerts, P., Matsuzawa, T., and Camerer, C.   Chimpanzee choice rates in competitive games match equilibrium game theory predictions. Scientific Reports, 4, 5182 doi:10.1038/srep05182 (2014).

3a. Lopata, J.   How Star Trek may show the emergence of human consciousness., November 18 (2014).

3b. Helbing, D., Yu, W., Opp, K-D., and Rauhut, H.   Conditions for the Emergence of Shared Norms in Populations with Incompatible Preferences. PLoS One, 9(8), e104207 doi:10.1371/ journal.pone. 0104207 (2014).

Here is an interesting collection of readings, which make sense in the context of the Martin paper. In Martin, the authors compare equilibrium expectations for Prisoner's Dilemma (PD) game [6] play with actual outcomes for both humans and chimps. It was found that when chimps play the game, the result is closer to the theoretical expectation than when humans play the game. This suggests that chimpanzee decision-making is more homogeneous than human decision-making, at least in the context of interactions that involve theory of mind. While the result is curious, there are two more recent items that might provide useful speculation about these outcomes.

In a recent article published in (3a), it is postulated that early human cognition (and perhaps cognition in the human-chimp common ancestor) resembled that of the Borg from Star Trek. It is argued that our ancestors were Borg-like in their ability exhibit little individuality across populations. In terms of the PD game, the theoretical equilibria often results from a convergence upon pure strategies. This may not so much a improvement upon an individual's ability to predict what their opponent will do next as the lack of heterogeneity in behavior over time. Thus, a species that exhibits much heterogeneity with respect to behavioral innovation (e.g. humans) would deviate from the theoretical expectation [7].

But does that mean the PD model is not really valuable in modeling human behavior? After all, the original formulation was modeled on human behavior. In addition, the model implicitly relies upon behavioral traits possibly unique to human cognition (such as theory of mind). In 3b, we can see that even though humans have a great diversity of preferences, sets of shared norms may emerge that serve to unify the behavioral outcomes of a population [8]. Despite our great individuality, some aspects of human culture can serve to reduce human heterogeneity.

[1] Michod, R.E.   Evolution of individuality during the transition from unicellular to multicellular life. PNAS, 104(S1), 8613-8618 (2007).

[2] Of course, there is plenty of debate regarding the role of group and multilevel selection in the evolutionary process. For more on the virtues of these types of selection, please see: Traulsen, A. and Nowak, M.A.   Evolution of cooperation by multilevel selection. PNAS, 103(29), 10952–10955 (2006) AND Goodnight, C.J.   Multilevel selection: the evolution of cooperation in non-kin groups. Population Ecology, 47, 3-12 (2005).

[3] The term "de-Darwinization" refers to the relaxation of selection for a given trait or level of selection. Please see: Godfrey-Smith, P.   Darwinian Populations and Natural Selection. Oxford University Press, New York (2009).

[4] Laland, K., Uller, T., Feldman, M., Sterelny, K., Muller, G.B., Moczek, A., Jablonka, E., Odling-Smee, J., Wray, G.A., Hoekstra, H.E., Futuyma, D.J., Lenski, R.E., Mackay, T.F.C., Schluter, D., and Strassmann, J.E.   Does evolutionary theory need a rethink? Nature, October 8 (2014).

[5] Richerson, P.J. and Boyd, R.   Natural Selection and Culture. BioScience, 34(7), 430-434 (1984) AND Bell, A.V.   Why cultural and genetic group selection are unequal partners in the evolution of human behavior. Communicative and Integrative Biology, 3(2), 159–161 (2010).

[6] Johnson, D.D.P., Stopka, P., and Bell, J.   Individual variation evades the Prisoner's Dilemma. BMC Evolutionary Biology, 2, 15 (2002).

[7] Herbert-Read, J.E., Krause, S., Morrell, L.J., Schaerf, T.M., Krause, J., and Ward, A.J.W.   The role of individuality in collective group movement. Proceedings of the Royal Society B, 280(1752), 1-8 (2013).

[8] Such norms, such as sharing, exhibit species differences in children. For an example, please see: Hamann, K., Warneken, F., Greenberg, J.R., and Tomasello, M.   Collaboration encourages equal sharing in children but not in chimpanzees. Nature 476, 328–331 (2011).

November 16, 2014

Thought (Memetic) Soup: November edition

This content is cross-posted to Tumbld Thoughts. Here are a few short observations on the state of the world and data, circa Summer 2014. Haven't gotten around to cross-posting these yet. The meta-theme is social disruption, evolutionary change, and economic dynamics, in spite of ideonational bias. These include Disruption du jour (I), Satire Makes it Doubly Skewed (II), and Ideonational Skew - Satire = Epistemic Closure? (III).

I. Disruption du jour

Is the idea of disruptive innovation a useful concept, or is it largely a misapplied buzzword. In the original definition of "creative destruction", Joseph Schumpeter described a process of innovation that resembled an avalanche or an earthquake. For example, most innovations do not reshape their respective industries, but a few key innovations (born out of creative ferment) do.

The modern notion of disruptive innovation does not make the distinction between the effects of innovation in different industries, nor are all so-called "disruptions" equally as valuable. Schumpeter's model of disruptive innovation resembles a power law, while the modern conception of disruptive innovation argues that transformative changes are ubiquitous. Here are some readings on the myth and controversies surrounding the concept:

Lepore, J.   The Disruption Machine. New Yorker, June 23 (2014).

* a critique of the "disruption" industry.

Bennett, D.   The Innovator's New Clothes: Is Disruption a Failed Model? Bloomberg Businessweek, June 18 (2014).

* perhaps Lepore is right -- disruption for disruption's sake is not a viable model of economic change.

Bennett, D.   Clayton Christensen Responds to New Yorker Takedown of 'Disruptive Innovation'. Bloomberg Businessweek, June 20 (2014).

* a rebuttal to the Lepore article from the modern "disruption" guru.

II. Satire Makes it Doubly Skewed

Two (intentionally) skewed views on Evolution [1, 2]: God does not do art, and monkeys still exist. Or something like that. Anyways, here is a sampling of creationism satire from Summer 2014.

[1] Pliny the In-Between   Theistic evolution. Evolving Perspectives blog, July (2014).

[2] Why There are Still Monkey (fake book in the Dummies series). Timothy McVeins Twitter post, June 20 (2014).

III. Ideonational Skew - Satire = Epistemic Closure?

Statistical conspiracy theory? Here is a link to John Williams' Shadowstats site and (appropriately) three readings [1-3] that critique the overall approach. For example, in one reading, it is suggested that the "shadow" in the Shadowstats name consists of an inappropriate modeling methodology.

[1] Aziz   The Trouble with Shadowstats. Azizonomics, June 1 (2013).

[2] Krugman, P.   Always Inflation Somewhere. Conscience of a Liberal blog, July 19 (2014).

[3] Hiltzik, M.   A new right-wing claim: Obama must be lying about inflation. The Economy Hub, Los Angeles Times, July 23 (2014).

November 6, 2014

The Top 100 Needles in a Haystack

A week or so ago, Nature News published a feature on the Top 100 (e.g. most-cited) articles of all time [1]. Interesting read, even if you don't agree with their methods or conclusions. Briefly, a Science Citation Index (SCI) was used to generate a list, the top 100 articles of which were considered the most highly-cited papers. Another more inclusive list was generated for comparison using Google Scholar. The outcomes were then evaluated.

The top papers-as-mountain peak analogy. COURTESY: Nature Publishing.

So which type of papers dominated the top of the list? As you might have guessed, papers that describe the details of a now-ubiquitous method dominate the top 100. Articles on methods such as single-step RNA isolation (#5) or density-functional thermochemistry (#8) have been cited in the neighborhood of 50-60,000 times because they provide a simple description of a method that has now become widespread. As it is considered good form to cite the source article for a given method, these are the top articles in this index. It may seem a bit disingenuous to count these articles as the most influential in science. For example, the Watson and Crick paper [3] describing the structure of DNA does not make the top 100. But, this methodology allows us to see the relative diffusion of such methods in the literature. Normalizing the number of citations by their respective age gives us a citation rate, which in turn allows us to estimate the velocity [4] of a given method through the scientific community.

Number 12 on the list is the paper that introduced the BLAST genome alignment method [2]. COURTESY: Nature Publishing.

For comparative purposes, the Nature News article also provides an alternate index, one compiled by Google Scholar. This index not only includes books, but deals with citations that are more conceptual in nature. For example, the top 100 citations of all time includes: Thomas Kuhn's "Structure of Scientific Revolutions" (#7), Claude Shannon's "Mathematical Theory of Communication" (#9), Rogers' "Diffusion of Innovations" (#17), "The Rat Brain in Stereotaxic Coordinates" (#23), and Zadeh's "Fuzzy Sets" (#29). Much like the SCI method, the Google Scholar method results in a long-tailed distribution, with a few papers far exceeding the rest of the literature in terms of citations. While it is less dominated by methods papers (and books), most of the top references on the list have nevertheless had a major influence on a number of fields. 

But what does all this mean for your recent publication? Will your recent paper on the mathematical structure of inter-neuronal conversion be worthy of one of these top spots at some point in the future? One interesting exercise might be to predict the future citation rates and diffusion velocities for recent papers and books (published within the last 10 years). While not at the very top of the list, these papers would demonstrate how the middle of the distribution lives. And for all of those papers with only a few citations even after years of being published, don't despair. It could be that your paper does not fit the criterion of a "top" paper (e.g. covers a series of clever but non-landmark studies), and so has not gained a high profile. Citation patterns are a curious thing.

[1] The relevant databases have limited this to the 20th and 21st centuries. Article citation: Van Noorden, R., Maher, B., Nuzzo, R.   The Top 100 Papers. Nature News, October 29 (2014).

[2] Full citation: Altschul, S.F., Gish, W., Miller, W., Myers, E.W., and Lipman, D.J.   Basic local search alignment tool. Journal of Molecular Biology, 215, 403-410 (1990).

[3] Watson, J.D. and Crick, F.H.C.   A Structure for Deoxyribose Nucleic Acid. Nature, 171, 737-738 (1953).

[4] Calculating method or publication "velocity" involves mapping the citation rate (per unit time) to a metric space or graph the represent the "shape" of the scientific community (disciplines, interest areas, etc). This is intentionally vague, as various community shapes are contingent on many factors.

October 31, 2014

Introducing the Evolution of Inequality Project

The study of income and resource inequality has been the academic topic du jour this year, highlighted by Thomas Piketty's economic history opus [1] and a special issue of Science [2]. There was even a paper on the 1% vs. the 99% of academic publishing [3] which argues that in terms of citations, the rich tend to get richer. But how do these patterns emerge and evolve? Is it a merely a statistical artifact, or a reflection of how complex, hierarchical societies tend to evolve? For example, we might assume that extreme inequality is maladaptive. But upon simulating a range of artificial societies of different population sizes, initial degrees of stratification, and behavioral features, we might find that extreme inequality tends to occur under specific conditions.

This is the motivating factor for my interest in the topic. Unlike many of the more well-versed approaches to the topic, an alternative view of inequality is a cross between statistical distributions, nuanced views of human sociobiology, and the aftermath of social change. While most economists have taken a materialist view of inequality, I feel that evolutionary perspectives would be helpful in teasing out questions of inequality's origins. This might involve data as diverse as historical data, ethnographic data, evolutionary modeling, and behavioral/neurophysiological data. In the end, we will be able to provide a conceptual alternative to the usual discussion at the intersection of behavior, biology, and social change. An inclusive, multidisciplinary approach is a core component of this project.

My research organization (Orthogonal Research) is trying to initiate work on a project called "The Evolution of Value and Inequality". This project is an attempt to understand the emergence of these social inequalities as a set of evolutionary and biobehavioral phenomena. While the argument can be made that an evolutionary perspective might be helpful in understanding unequal allocations of resources, it is sorely lacking in the general discussion. The broadness of the initiative is necessary to make the connection between the pure inferential approach of evolutionary science coherent public policy outcomes. An initial grant application to the Washington Center for Equitable Growth (submitted January 2014) did not get funded, one reason being that the idea needed more conceptual and empirical fleshing out.

So upon doing some more conceptual refinement, I just finished submitting a second version of this proposal to the Institute for New Economic Thinking (INET). While I will not get into the technical details here, the basic idea is to construct adaptive computational models that mimic a social hierarchy (so-called hierarchical network models). Each node of these directed graphs are informed in their behavior by neuroimaging and other physiological sources of data on human behavior.

The project is focused around evolutionary models of social change (social and cultural change), the underlying assumptions of which can be verified by the collection of biobehavioral data (e.g. neuroimaging experiments). The empirical component is meant to test assumptions of individual and social behavior, and serves as an alternative to the rational expectations assumption that dominates much of conventional economics. However, it also refines many of the model-free findings that
characterize behavioral economics [4].

The evolutionary aspects of this work are also quite interesting. Because we are bridging the short-term (behavioral) and the longer-term (social evolution), there are at least three forms of adaptation: a social learning mechanism, a cultural evolutionary form of selection, and a neurophysiological imperative that satisfies various material, social, and existential needs of an individual. This gives us a fitness and selection criterion that is tangentially related to reproductive success. Subsequent evolutionary algorithms and simulations may bear out the evolutionary dynamics of value construction and social stratification.

Another contribution of this project is to link the statistical aspects of inequality with an evolutionary and demographic framework. The oft-referenced phrase the "1%" or the "0.01%" has its roots in an exponential (non-normal) statistical distribution called the power law. Power laws of various size tend to describe observed income distributions in many different types of society. As inequality increases sharply in a single society, or as different degrees of inequality are observed in different contexts, the power law and in conjunction with various stable states can be used as selection criteria.

Schematic of expected results. Both the C and L parameters refer to operations on intra- and bi-level hierarchical networks dynamics, respectively.

As proposed in the first part of this post, the resulting evolutionary algorithms and experimental inquiries provide us with a possibility space for given outcomes. Given a set of initial conditions, we can observe the tendencies of inequality of resource allocation. If there are common outcomes relative to a number of different initial conditions, this could tell us something cross-cultural and fundamental about the nature of inequality. Ultimately, the outcomes of this project could help to identify and predict opportunities to head off crises and as an architecture for achieving sustainable economic growth.

[1] Piketty, T.   Capital in the 21rst Century. Harvard University Press (2014).

[2] Citation for the special issue: Science, 344, May 23 (2014). One article with particular relevance to social evolution is: Pringle, H.    The Ancient Roots of the 1%. Science, 344, 822-825 (2014).

[3] Ioannidis JPA, Boyack KW, Klavans R.   Estimates of the Continuously Publishing Core in the Scientific Workforce. PLoS One, 9(7), e101698. doi:10.1371/journal.pone.0101698 (2014).

[4] Camerer, C.F. and Loewenstein, G.   Behavioral Economics: past, present, future. In "Advances in Behavioral Economics". C.F. Camerer, G. Loewenstein, and M. Rabin eds. Chapter 1. Russell Sage Foundation (2004).

Behavioral Economics Reading List. Russell Sage Foundation blog, March 23 (2012).

October 26, 2014

C. elegans as an Evolutionary Model

In the past year, I have been starting to use the nematode Caenorhabditis elegans (roundworm) as a model organism. Not only have I helped to establish the DevoWorm project, I am also starting to engage with C. elegans in a wet-lab setting. As a consequence, I am learning about multiple facets of C. elegans biology. C. elegans is a well-established model organism, having well-characterized neural and developmental systems. The nervous system contains just 302 cells, with a full accounting of the connectome (synaptic connections) [1]. The developmental system is also well-characterized, with a lineage tree [2] having been worked out for the entire organism. While a lineage tree relies upon deterministic mechanisms (and thus cannot be applied to organisms such as Mammals), it does provide us with a clear accounting of cell differentiation and organ formation during development. Thus, C. elegans is a tractable model for whole-organism investigations (Figure 1).

Figure 1. Anatomy of the adult hermaphrodite. COURTESY: WormAtlas.

But what about evolution? At first pass, it seems as though asking evolutionary questions is not a tractable feature of roundworm biology. Nevertheless, we can use this worm to answer several outstanding questions in evolution [4]. I will use information from a recent review by Jeremy Gray and Asher Cutter [5] to discuss these potential research advances (Figure 2). The actual future applications of C. elegans as an evolutionary model might turn out to investigate other issues. As it turns out, the roundworm provides a happy medium between more traditional models of experimental evolution (microbes) and complex organisms with long generation times (humans). While C. elegans have a relatively short generation time (~50 hours), they also have complex phenotypes with organs.

Figure 2. The life cycle and means of experimental manipulation for evolution experiments. COURTESY: Figure 1 in [5].

The most common means of experimental evolution proposed in [5] is the mutation accumulation (MA) approach. MA may also serve as a weak factor in determining life-history traits in a species [6]. In experimental evolution, the MA approach allows us to observe the role of mutational variation in evolution. One way to apply this method might be to manipulate a single gene (using directed mutagenesis, gene editing, or RNAi -- see Figure 2) and then place it in a genetic background. Rather than waiting for a series of mutations to emerge in a population, mutation is induced to maximize the variation upon which evolution can act upon [7].

Another means of experimental evolution discussed in [5] is co-evolution between worms and pathogens. This can done by culturing worms in ecological context over several generations. One prediction involves the evolution of tradeoffs observed in already co-evolved relationships such as C. elegans growth rate and pathogenic resistance [8]. A secondary means of understanding the ecology of evolution involves introducing environmental fragmentation through introducing spatial variation (physical barriers or agar gradients) on a culture dish. This can produce to genetic bottlenecks and other effects related to population structure and neutral processes.

A third strategy discussed in [5] involves examining different reproductive strategies and degrees of adaptability between species of Caenorhabditis. The latter topic might include a better understanding of how the degeneracy [9] of neuronal and genetic circuits that lead to observable behaviors and phenotypes evolves. Yet there is also great potential for C. elegans to be used as an eco-evo-devo [10] model which integrates to response of environmental stimuli by cell and molecular mechanisms of development over evolutionary time (Figure 3). While I do not have plans on establishing my own C. elegans experimental evolution program in the near future, stay tuned.

Figure 3. A fledgling eco-evo-devo approach to C. elegans.

[1] Jarrell, T.A., Wang, Y., Bloniarz, A.E., Brittin, C.A., Xu, M., Thomson, J.N., Albertson, D.G. Hall, D.H., and Emmons, S.W.   The Connectome of a Decision-Making Neural Network. Science, 337, 437-444 (2012).

Please also see The Connectome Project website.

[2] Sulston, J.E., Schierenberg, E., White, J.G., and Thomson, J.N.   The Embryonic Cell Lineage of the Nematode Caenorhabditis elegans. Developmental Biology, 100, 64-119 (1983).

[3] Jovelin, R., Dey, A., Cutter, A.D.   Fifteen Years of Evolutionary Genomics in Caenorhabditis elegans. eLS, doi:10.1002/9780470015902.a0022897 (2013).

[4] For a review of Caenorhabditis phylogeny and evolutionary biology, please see: Fitch, D.H.A. and Thomas, W.K.   Evolution. In "C. elegans II", Chapter 29. Cold Spring Harbor Laboratory, Woods Hole, MA (1997).

[5] Gray, J.C. and Cutter, A.D.   Mainstreaming Caenorhabditis elegans in experimental evolution. Proceedings of the Royal Society B, 281, 20133055 (2014).

[6] Danko, M.J., Kozlowski, J., Vaupel, J.W., and Baudisch, A.   Mutation Accumulation May Be a Minor Force in Shaping Life History Traits. PLoS One,  7(4), e34146 (2011).

[7] Thompson, O., Edgley, M., Strasbourger, P., Flibotte, S., Ewing, B., Adair, R., Au, V., Chaudhry, I., Fernando, L., Hutter, H., Kieffer, A., Lau, J., Lee, N., Miller, A., Raymant, G., Shen, B., Shendure, J., Taylor, J., Turner, E.H., Hillier, L.W., Moerman, D.G., and Waterston, R.H.   The million mutation project: a new approach to genetics in Caenorhabditis elegans. Genome Research, 23(10), 1749-1762 (2013).

[8] Schulte, R.D., Makus, C., Hasert, B., Michiels, N.K., and Schulenburg, H.   Multiple reciprocal adaptations and rapid genetic change upon experimental coevolution of an animal host and its microbial parasite. PNAS USA, 107, 7359 –7364 (2010).

[9] Degeneracy involves structurally different elements (such as functional neuronal networks) that converge upon the same output. An example of this within C. elegans: Trojanowski, N.F., Padovan-Merhar, O., Raizen, D.M. and Fang-Yen, C.   Neural and genetic degeneracy underlies Caenorhabditis elegans feeding behavior. Journal of Neurophysiology, 112, 951-961 (2014).

[10] Abouheif, E., Fave, M.J., Ibarraran-Viniegra, A.S., Lesoway, M.P., Rafiqi, A.M., and Rajakumar, R.   Eco-evo-devo: the time has come. Advances in Experimental Medicine and Biology, 781, 107-125 (2014).