Deep Reading Brings New Things to Life (Science)

Here is an interesting Twitter thread from Jacquelyn Gill on 'deep reading':

The basic idea is that exploring older literature can lead to new insights, which in turn lead to new research directions. The new research of our era tends to focus on the most relevant and cutting-edge literature [1]. This recency bias excludes many similarly relevant articles, including articles that perhaps inspired the more recent citations to begin with [2]. 

I have my own list of deep reads that have influenced some of my research in a similar fashion. These references can be either foundational or so-called "sleeping beauties" [3]. Regardless, I am doing my part to maintain connectivity [4] amongst academic citation networks:

1) Woodger, J.H. The Axiomatic Method in Biology. 1937.

An argument for biological rules, an influence on cladistics (developed in the 1960s), and a natural bridge to geometric approaches to data analysis and modeling. While there is a strong argument to be made against the axiomatic approach [5], this directly inspired much of my thinking in the biological modeling area. 

2) Davis R.L., Weintraub H., and Lassar A.B. Expression of a single transfected cDNA converts fibroblasts to myoblasts. Cell 51, 987–1000. 1987.

This was the first proof-of-concept for direct cellular reprogramming, and predates the late 2000's Nobel-winning work in stem cells by decades. In this case, a single transcription factor (MyoD) was used to convert a cell from one phenotype to another without a strict regard for function. More generally, this paper helped inspired my thinking in the area of cellular reprogramming to go beyond a biological optimization or algorithmic approach [6].

3) Ashby, W.R. Design for a Brain. 1960.

"Design for a Brain" serves as a stand-in for the entirely of Ashby's bibliography, but this is the best example of how Ashby successfully merged explanations of adaptive behavior [7] with systems models (cybernetics). In fact, Ashby originally coined the phrase "Intelligence Augmentation" [8]. I first discovered Ashby's work while working in the area of Augmented Cognition, and has been more generally useful as inspiration for complex systems thinking.

4) Schroeder, M. Fractals, Chaos, and Power Laws: minutes from an infinite paradise. W.H. Freeman. 1991. AND Kauffman, S. At Home in the Universe: search for the laws of self-organization and complexity. Oxford University Press. 1995.

Not so much a couple of sleeping beauty as easy reading technical reference guides for all things complexity theory.

5) Bourdieu, P. Outline of a Theory of Practice. Cambridge University Press. 1977 AND Alexander, C., Ishikawa, S., and Silverstein, M. A Pattern Language: towns, buildings, construction. Oxford

University Press. 1977.

This is a bonus, not because the references are particularly obscure or even from the same academic field, but because they partially influenced my own view of cultural evolution. This is yet another piece of advice to young researchers: take things that appear to be disparate on their surface and incorporate them into your mental model. If nothing else, you will gain valuable skills in intellectual synthesis.

UPDATE (11/17):

Here is another example of old (classic, not outdated) work influencing new scholarship.

NOTES:

[1] Evans, J.A. (2008). Electronic Publication and the Narrowing of Science and Scholarship. Science, 321(5887), 395-399 AND Scheffer, M. (2014). The forgotten half of scientific thinking. PNAS, 111(17), 6119.

[2] related topics discussed on this blog include distributions of citation ages and most-cited papers.

[3] van Raan, A.F.J. (2004). Sleeping Beauties in Science. Scientometrics, 59(3), 467–472.

[4] Editors (2010). On citing well. Nature Chemical Biology, 6, 79.

[5] For the semantic approach (which had been influential to my more recent work), please see: Lloyd, E.A. (1994). The Structure and Confirmation of Evolutionary Theory. Princeton University Press, Princeton, NJ.

[6] Ronquist, S. et.al (2017). Algorithm for cellular reprogramming. PNAS, 114(45), 11832–11837.

[7] Sterling, P. and Eyer, J. (1988). Allostasis: A new paradigm to explain arousal pathology. In "Handbook of life stress, cognition, and health". Fisher, S. and Reason, J.T. eds. Wiley, New York. 

[8] Ashby, W.R. (1956). An Introduction to Cybernetics. Springer, Berlin.

New Papers, Old Papers, and Re-convolved Concepts, November edition

I have been busy the past several months fleshing out new ideas and finishing up older ones. The first paper profiled here is "Cellular decision-making bias: the missing ingredient in cell functional diversity", something I published on arXiv [1] last month. This paper is a computational-oriented derivative of the paper "Defining phenotypic respecification diversity using multiple cell lines and reprogramming regimens", published earlier this year in Stem Cells and Development [2].

In [2], it was demonstrated that a series of different cell lines of the same type (e.g. fibroblast) exhibit great variability (many-fold differences) in terms of their direct cellular reprogramming efficiency. The efficiency of this process was measured using phenotypic (e.g. immunocytochemical) assays. This may or may not be due to the underlying genomic processes. Using a limited set of assays analyzed by means of differential gene expression, no smoking gun was found. While we did not investigate candidate epigenetic markers, the phenotypic trend was nevertheless consistent for both human and mouse cells reprogrammed to both generic muscle fiber and generic dopaminergic neurons [3].

The data collected and analyzed here also sets up a series of computational investigations using a method derived from Signal Detection Theory (SDT) and other signal-to-noise characterization methods [4]. SDT is generally used to understand cognitive decision-making in humans and animals. However, decision-making theory has also been used to explain outcomes at the cellular and molecular level, particularly switch-like processes [5]. Using the standard SDT as inspiration, I propose in [1] that cellular and molecular processes can be characterized and analyzed using a technique called cellular SDT.

Major collaborator on the Stem Cells and Development paper [2]: Dr. Steven Suhr, Michigan State University. 

Cellular SDT can uncover something called decision-making bias, which is hypothesized to occur during the conversion of cells from one phenotype to another [3]. In this case, the term bias refers to the magnitude of difference in conversion efficiency for the same cell line given two distinct stimuli. The overarching assumption is that differences observed across different small-scale stimuli (e.g. forced transcription factor activity) can be characterized systematically within and between specific cell types and lines.

My talk to the BEACON Center in May 2013. The first part (YouTube video) focused on modeling diversity in cellular reprogramming (an early version of cellular decision-making bias).

Here is the abstract of the paper. Associated code (on Github) can be found here:

"Cell functional diversity is a significant determinant on how biological processes unfold. Most accounts of diversity involve a search for sequence or expression differences. Perhaps there are more subtle mechanisms at work. Using the metaphor of information processing and decision-making might provide a clearer view of these subtleties. Understanding adaptive and transformative processes (such as cellular reprogramming) as a series of simple decisions allows us to use a technique called cellular signal detection theory (cellular SDT) to detect potential bias in mechanisms that favor one outcome over another. We can apply method of detecting cellular reprogramming bias to cellular reprogramming and other complex molecular processes. To demonstrate the scope of this method, we will critically examine differences between cell phenotypes reprogrammed to muscle fiber and neuron phenotypes. In cases where the signature of phenotypic bias is cryptic, signatures of genomic bias (pre-existing and induced) may provide an alternative. The examination of these alternates will be explored using data from a series of fibroblast cell lines before cellular reprogramming (pre-existing) and differences between fractions of cellular RNA for individual genes after drug treatment (induced). In conclusion, the usefulness and limitations of this method and associated analogies will be discussed."

The second paper profiled here is called "A Semi-automated Peer-review System", a short paper I published on the arXiv earlier this month [6]. The idea of an automated peer review system came to me after preparing a blog post [7] and reading a paper on the most common degree of novelty found among highly influential scientific papers [8]. The paper provides an outline of a human-assisted adaptive algorithm that detects fraud in a set of scientific papers without also filtering out innovative but highly-novel work. As in the case in [1], the approach was based on signal detection theory (SDT). In this case, however, a more conventional application (e.g. standard ROC curves) is used to minimize the number of truly low quality and fraudulent manuscripts while maintaining diversity and novelty in the scientific literature.

Here is the abstract and here is the associated code (mostly pseudo-code) on Github:

"A semi-supervised model of peer review is introduced that is intended to overcome the bias and incompleteness of traditional peer review. Traditional approaches are reliant on human biases, while consensus decision-making is constrained by sparse information. Here, the architecture for one potential improvement (a semi-supervised, human-assisted classifier) to the traditional approach will be introduced and evaluated. To evaluate the potential advantages of such a system, hypothetical receiver operating characteristic (ROC) curves for both approaches will be assessed. This will provide more specific indications of how automation would be beneficial in the manuscript evaluation process. In conclusion, the implications for such a system on measurements of scientific impact and improving the quality of open submission repositories will be discussed". 

Finally, I am giving a presentation at the Network Frontiers Workshop at Northwestern University's NICO Institute on the 4th of December. The title of the talk is "From Switches to Convolution to Tangled Webs: evolving sub-optimal, subtle biological mechanisms". The work is an extension of my arXiv paper from 2011 [9] on Biological Rube Goldberg Machines (RGMs), something I also refer to as a convolution architecture. Here is the abstract and here is the associated code on Github:

"One way to understand complexity in biological networks is to isolate simple motifs like switches and bi-fans. However, this does not fully capture the outcomes of evolutionary processes. In this talk, I will introduce a class of process model called convolution architectures. These models demonstrate bricolage and ad-hoc formation of new mechanisms atop existing complexity. Unlike simple motifs (e.g. straightforward mechanisms), these models are intended to demonstrate how evolution can produce complex processes that operate in a sub-optimal fashion. The concept of convolution architectures can be extended to complex network topologies. Simple convolution architectures with evolutionary constraints and subject to natural selection can produce step lengths that deviate from optimal expectation. When convolution architectures are represented as components of bidirectional complex network topologies, these circuitous paths should become “spaghetti-fied”, as they are not explicitly constrained by inputs and outputs. This may also allow for itinerant and cyclic self-regulation resembling chaotic dynamics. The use of complex network topologies also allows us to better understand how higher-level constraints (e.g. hub formation, modularity, preferential attachment) affect the evolution of sub-optimality and subtlety. Such embedded convolution architectures are also useful for modeling physiological, economic, and social complexity". 

And last but not least, a new preprint server has come online called BioRxiv. BioRxiv (administered by Cold Spring Harbor Laboratory) accepts manuscripts from a number of biological disciplines, from Bioinformatics to Molecular Biology to Zoology. I kicked things off in the Zoology category with an older manuscript (originally presented at a conference in 2006) entitled "Filling up the Tree: considering the self-organization of avian roosting behavior" [10]. However, for more theoretical and interdisciplinary work such as the paper in [11], I still plan on using arXiv.

NOTES:

[1] Alicea, B.   Cellular decision-making bias: the missing ingredient in cell functional diversity. arXiv repository, arXiv: 1310:8268 [q-bio.QM] (2013).

[2] Alicea, B., Murthy, S., Keaton, S.A., Cobbett, P., Cibelli, J.B., and Suhr, S.T.   Defining phenotypic
respecification diversity using multiple cell lines and reprogramming regimens. Stem Cells and Development, 22(19), 2641-2654 (2013).

[3]  In this example, conversion refers to direct cellular reprogramming technique (e.g. the creation of iPS cells) that result in the creation of induced neural cells (iNCs) and induced skeletal muscle cells (iSMCs). However, conversion could also refer to carcinogenesis or developmental processes.

Figure 1 from Alicea et.al (2013). Frames A-D, immunocytochemical characterization of iNCs and iSMCs. Frames E-H, diversity in reprogramming efficiency for a range of cell lines.

[4] Schultz, S.R.   Signal-to-noise ratio in neuroscience. Scholarpedia, 2(6), 2046 (2007).

[5] Balazsi, G., van Oudenaarden, A., and Collins, J.J.   Cellular Decision-Making and Biological Noise: From Microbes to Mammals. Cell, 144(6), 910–925 (2011). 

[6] Alicea, B.   A Semi-automated Peer-review System. arXiv: 1311.2504 [cs.DL, cs.HC, cs.SI, physics.soc-ph] (2013).

[7] Alicea, B.   The Novelty-Consensus Dampening.   Synthetic Daisies blog, October 22 (2013). 

[8] Uzzi, B., Mukherjee, S., Stringer, M., and Jones, B.   Atypical Combinations and Scientific Impact. Science, 342, 468-472 (2013).

[9] Alicea,  B.   The ‘Machinery’ of  Biocomplexity:  understanding  non-optimal  architectures  in biological systems. arXiv repository, arXiv: 1104.3559 [nlin.AO, q-bio.QM, q-bio.PE] (2011).

[10] Alicea, B.   Filling up the Tree: considering the self-organization of avian roosting behavior. bioRxiv, doi:10.1101/000349 (2013).

[11] Alicea, B.   The Emergence of Animal Social Complexity: theoretical and biobehavioral evidence. arXiv repository, arxiv:1309.7990 [q-bio.PR, q-bio.NC] (2013).

Evolution, Variation, Development, and Strains of Artificial Life in the Reading Queue

This content is cross-posted to my micro-blog, Tumbld Thoughts. Many new papers on adaptation and evolution, plus a call for conference participation in Artifical Life XIV.

New Readings on Short- and Long-term Evolution from the Reading Queue

Here are a few new papers on experimental evolution. The first is a paper from Jeffrey Barrick and Rich Lenski [1], who utilize the long-term evolution experiment to look at genome dynamics during bacterial evolution [2]. The first figure shows the types of mutations observed during evolution (occurring on the scale of 10³ generations). The second demonstrates the signatures of optimization, innovation, and epistasis in evolutionary change. Interestingly, a genomic analysis of bacterial populations from the same project suggests that adaptation proceeds without reaching so-called fitness peaks (which is predicted by theory to limit the fitness advantage of a given genotype).

The second paper is from Ted Garland, and involves using artificial selection [4] in mice to find the limits of evolution (or evolvability) over 10-100 generations [5, 6]. A wheel running task is used to assess physical performance. The first figure shows baseline performance, maximum evolved performance, and post-peak performance given genetic (G), environmental (E), and GxE sources of variation. The second figure shows differences in wheel running performance between male and female mice over 30 generations. In this case, behavioral analysis  reveals distinct limits to advantages gained from artificial selection (which are not always due to adaptation). 

New Readings on Human Variation from the Reading Queue

Here are four new papers on human genomic variation. The first [7] is a review of genome mosaicism, or variation across cells in the same human body. Mosaicism results from errors in either chromosome segregation during mitosis or DNA replication. In neurons from the frontal cortex [8], mosaicism is responsible for variation in chromosomal complements and copy number variants (CNVs). This variation comes in the form of aneuploidies, retrotransposons, and large-scale CNV differences (in 13-41% of neurons sampled). 

In [9], variation in chromatin states across the genome is explored. One finding suggests that variable regions are enriched in SNPs relative to nonvariable regions, which may be due to negative selection. The expression of heterozygous SNPs with allele-specific signals are highest for active marks. These is also variation in methylation switches (active/repressed or active/weakly active states) which results in enhancer and core promoter-specific states.

Finally, functional genomic elements can be more explicitly linked to chromatin signatures. This was done in [10] by finding the cis-regulatory variants that most affect chromatin states. In this study, five post-transcriptional modifiers and three transcription factors were used to show these trends across 14 individuals. It was found that allele-specific patterns of association (between genomic function and chromatin regulation) exist.

Calls for Artificial Life

If you enjoy creating artificial life, and want to write an academic paper about it (8 page, single-spaced limit, IEEE format), then you will want to submit your work to the Artificial Life 14 conference, being held Summer 2014 in NYC. Submission deadline (full papers) is March 31.

Topics include: bio-inspired robotics, cellular automata and artificial chemistries, synthetic life, embodied systems, collective behavioral dynamics, ecological/social/evolutionary dynamics, and the art and philosophy of Artificial Life. There is a separate call for workshops/tutorials (due January 15) and a Science Visualization competition (applications due February 1).

And, last but not least, some new Developmental Biology.....

Last but not least, here is a nice article by Carl Zimmer [11] summarizing the cutting-edge work being done on understanding the potential role of senescent cells in embryonic development. The excellent picture shows a mouse embryo (E15) with the areas of senescent cells stained in blue.

NOTES:

[1] Lenski's long-term evolution experiment was recently profiled in Science. Listen to this podcast for more: Crespi, S.   Podcast Interview: Richard Lenski. Science Express, November 14 (2013).

[2] Barrick, J.E. and Lenski, R.E.   Genome dynamics during experimental evolution. Nature Reviews Genetics, 14 827-839 (2013).

[3] Wiser, M.J., Ribeck, N., and Lenski, R.E.   Long-Term Dynamics of Adaptation in Asexual Populations. Science, DOI: 10.1126/science. 1243357.

[4] Postma, E., Visser, J., Van Noordwijk, A.J.   Strong artificial selection in the wild results in predicted small evolutionary change. Journal of Evolutionary Biology, 20, 1823–1832 (2007). 

[5] Careau, V., Wolak, M.E., Carter, P.A., and Garland, T.   Limits to Behavioral Evolution: the quantitative genetics of a complex trait under directional selection. Evolution, 67(11), 3102-3119 (2013).

[6] Barton, N. and Partridge, L.   Limits to natural selection. BioEssays, 22, 1075-1084 (2000).

For a short primer on the concept, please see this primer from Understanding Evolution.

[7] Lupski, J.R. et.al  One Human, Multiple Genomes: Genome Mosaicism. Science, 341, 358-359 (2013).

[8] McConnell, M.J. et.al   Mosaic Copy Number Variation in Human Neurons. Science, 342, 631-637 (2013).

[9] Kasowski, M. et.al  Extensive Variation in Chromatin States Across Humans. 750-752. Science, 342, 750 (2013).

[10] Kilpinen, H. et.al   Chromatin Structure, and Transcription Coordinated Effects of Sequence Variation on DNA Binding. Science, 342, 744-747 (2013).

[11] Zimmer, C.   Signs of Aging, Even in the Embryo. NYT Science, November 21 (2013).

Short Threads of Reading Queue

Here are some academic papers, articles, and blog posts I have put into my reading queue over the past few weeks that I have found interesting and/or comment-worthy. I have organized them into threads (e.g. streams of consciousness) here:

Short thread on cell biology and genomics:

[1] Xie, J. et.al   Autocrine signaling based selection of combinatorial antibodies that transdifferentiate human stem cells. PNAS, doi:10.1073/pnas.1306263110 (2013).

[2] Williams, R.B.H. et.al   The influence of genetic variation on gene expression. Genome Research, 17, 1707-1716 (2007).

In [1], the researchers use a combination of receptor antibodies to reprogram a cell's fate. Yet more evidence that cellular reprogramming is not only possible, but involves more than just a few transcription factors or a spontaneous transformation. The science in [2] is a pre-RNA-seq study on the effects of standing genome variation on steady-state gene expression. A good early review, although there is now more current/specific work available.

Short thread on economics, markets, and technology:

[1] Yglesias, M.   Who gets rich when robots take our jobs. Moneybox blog, May 13 (2013).

Mr. Spacely from "The Jetsons". He's rich and George Jetson is not.

[2] Falk, A. and Szech, N.   Morals and Markets. Science, 707, 340 (2013).

After reading [1], I come away with the impression that the only thing that can be economically gained from automation is a bolstering of the arbitrary claim (e.g. Russian roulette) to genius (e.g. even patent trolling qualifies). Apparently, it is more relevant (and fleeting) than ever. This is part of a trend that has lead to productivity gains of the last 40 years becoming locked up in corporations and/or an executive elite. Again, automation has helped this trend along, although automation does not always result in this outcome.

In [2], a curious finding is reported. If you are part of a market, you are more likely to let a mouse die for a lower amount of money. A novel addition to the experimental moral philosophy field. Not quite sure if this is an exercise in mutually-assured moral behavior (e.g. bystander effect), or a call to make judgments about economic value in isolation. Is there more than meets the eye to this simple set of experiments? As an aside, how does this relate to the psychology of auctions?

Short thread on subjectivity in the brain:

[1] Wittmann, M. et.al   The neural substrates of subjective time dilation. Frontiers in Human Neuroscience, doi:10.3389/neuro.09.002.2010 (2010).

[2] Schurger, A. et.al   Reproducibility distinguishes conscious from nonconscious neural representations. Science, 327, 97 (2010).

Apparently, 2010 was a good year for investigating subjectivity in the brain. How do we measure engagement with a piece of art or the practice of culture? In [1], changes in activity patterns among the "cognitive control" and "default activity" brain networks mediate subjective responses to visual motion. In [2], neural activity related to conscious, neural correlates of subjectivity must be both of a certain duration and intensity as well as being reproducible. While subjective experiences can be transient and unique, their neural correlates are not.

Happy 50th birthday, Chaos theory!

[1] Arbesman, S.   The Fiftieth Anniversary of Chaos. Social Dimension blog, May 17 (2013).

[2] Lorenz, E.N.   Deterministic Nonperiodic Flow. Journal of Atmospheric Science, 20, 130-141 (1963).

[3] Motter, A.E. and Campbell, D.K.   Chaos at Fifty. Physics Today, May, 27 (2013).

This feature got a pretty decent response on my micro-blog, Tumbld Thoughts: happy 50th birthday to the study of chaos [1]. A worldview first proposed (in formal fashion) by Edward Lorenz in a landmark paper on weather prediction called “Deterministic Nonperiodic Flow” [2]. Later, the field would grow to encompass analytical strategies such as nonperiodic attractors, bifurcation maps, and fractals.

As a new way to describe physical phenomena and complex systems with a high degree of nonlinearity and subtle unpredictabilities (e.g. the butterfly effect), chaos shattered the notion of a clockwork universe [3]. As a paradigm shifting concept, chaos theory has the potential to enrich all areas of science [4].

Image on left is from [1], and image at the right is from [3]. For the latest work in the field, check out the journal “Chaos: an interdisciplinary journal of nonlinear science”.

For examples from brain science, see the following two articles and book:

* Robson, D.   Disorderly genius: How chaos drives the brain. New Scientist, June 29 (2009). YouTube video.

* Kitzbichler, M.G., Smith, M.L., Christensen, S.R., Bullmore, E.   Broadband Criticality of Human Brain Network Synchronization. PLoS Computational Biology, 5(3), e1000314 (2009).

* Freeman, W.J.   Neurodynamics: an exploration in mesoscopic brain dynamics. Springer, Berlin (2006).

Intriguing evolution stuff:

Zimmer, C.   Enlisting a virtual pack, to study canine minds. New York Times, April 22 (2013).

The Dognition website.

This is a story about Dr. Hare, the Anthropologist (the study of humans) interested in canine cognition. Can we throw any more species in there? Oh yes -- apparently dogs are more intelligent than their wolf wild-type cousins (determined by something called the "pointing test"). So to make this assessment more scientific, Dr. Hare came up with a test for dog intelligence. He also founded a company called Dognition, which is collecting data from dogs worldwide. But there's no such thing as a Dog IQ just yet. It will be interesting to see how intelligence corresponds with breed and degree of artificial selection for specific traits.

Evolutionary "gut check":

Burger, O. et.al   Human mortality improvement in evolutionary context. PNAS, doi:10.1073/ pnas.1215627109 (2013).

This is a paper that I could not quite figure out. My gut says that something is not quite right/being accounted for here. Are they using ethnographically-observed hunter gatherer populations to derive an evolutionary baseline? If so, can they truly demonstrate that these populations actually represent such a baseline? Also, it seems to me that increases in life expectancy may involve the elimination of early mortality (due to warfare, violence, and disease) rather than a biological or cultural adaptation (particularly one on the order of those that distinguish between sister taxa, as the one that distinguishes human hunter-gatherers and chimps).

...and, finally, actual robots!

ICRA 2013 Conference website. Held in Karlsruhe, Germany, and sponsored by IEEE.

Erico Guizzo reports for IEEE Spectrum from ICRA (Robotics Conference), and brings us (among many other interesting things) a feature on Entropica:

LEFT: Screenshots of Entropica configurations (social network interactions and a pole balancing task). RIGHT: real-world (e.g. Primate) behaviors (termite dipping/tool use and stock market trading).

Wissner-Gross, A.D. and Freer, C.E.   Causal Entropic Forces. Physical Review Letters, 110, 168702 (2013).

Hewitt, J.   The emergence of complex behaviors through causal entropic forces. Phys.org, April 22 (2013).

Using a robotic model, it can be demonstrated that general intelligence (in the form of causal generalization) may be amplified or otherwise result from entropy maximization. This is related to work done on ant trails, showing that they conform to Fermat's principle of least time.

Lecture to BEACON Center, Michigan State

On May 17th (Friday) at 3:30pm (EST), I will be giving a lecture entitled "Adventures in Quasi-Evolution" [1] to the BEACON Center [2]. The audience is Thrust Group 1 (Genomes, Networks, and Evolvability).

The first half of my talk will be on computational models of cellular reprogramming (e.g. evolutionary modulus, or the engineering on the remnants of evolutionary and developmental variation). The second half is some emerging work I am doing on the cultural evolution of economic value (e.g. evolutionary through the looking glass, or how evolutionary models [3] may explain current economic puzzles).

NOTES:

[1] "Adventures in Quasi-Evolution". Figshare, doi:10.6084/m9.figshare.701463 (2013). I define quasi-evolution as "changes over time not due to reproductive fitness or generational inheritance".

[2] For those who are unfamiliar, the BEACON Center (an NSF-funded center) is a multi-disciplinary, multi-University group interested in the intersection of biology and engineering with relevance to evolution. Catch the lecture at one of these locations:

Michigan State University: Biomedical and Physical Sciences Building, Room 1441 (BEACON seminar room), 3:30pm.

North Carolina A and T University: McNair Hall, Lecture Room 4, 3:30pm.

University of Idaho: Life Sciences South (LSS), Room 144, 12:30pm

University of Texas, Austin: Service Building (SER), Room 321H, 2:30pm.

University of Washington: UW Hutchinson Cancer Center, Building PAA Room 023D, 12:30pm

[3] The cultural evolutionary models used in my research are called Contextual Geometric Structures (CGSs). Contextual Geometric Structures (CGS) will never play chess well, or perhaps at all. They will never beat Ken Jennings on Jeopardy. That's not the point. They exist as soft or fuzzy (e.g. possibilistic, non-transitive) classifiers that capture (or at least reproduce) the structural features of cultural behavior. This is quite different from the dual inheritance models that are common in studies that focus on the geneaology of traits.

The "structure" of culture has been observed and pondered by many cultural anthropologists, from Claude Levi-Strauss to Pierre Bourdieu. Bourdieu used a construct called the "habitus" to characterize the relationship between individuals in a single generation and cultural structures that exist across multiple generations. CGSs provide a measure of computational precision (e.g. a kernel function)  to this and other conceptions of "cultural logic".

As a quasi-evolutionary phenomenon, CGSs are designed to capture neither the dual inheritance of genes and culture nor the connectionism of a traditional cognitive model. The outcomes of CGS agents are not geared towards optimal decision-making. Rather, they classify natural phenomena according to a set of discrete oppositions (or categories).

Some of these are based on premises (e.g. historically-determined preferences), while others are based on biological features of the organism. The CGS agents then use the classificatory state of other agents as a cue to either follow (conform) or disperse (dissent). What results is a form of social learning that can be used to update (or obliterate) the space between the lower-level categories.

CGS simulations can be run either in an liquid-like simulation, or independently. In the original conception, which maps CGSs to geographic and other spatial phenomena, a hybrid model was proposed. In experiments geared towards understanding the social construction of economic value, populations of CGS kernels and their agents are static with respect to spatial position. However, they exchange items and attach value to those items based on repeated interaction.

For more information, please see:
Alicea, B.  Contextual Geometric Structures: modeling the fundamental components of cultural behavior. Proceedings of Artificial Life, 13, 147-154 (2012).

Replication, Model Organisms, and the Role of Evolutionary Signatures

The following slides and commentary focus on an open problem that involves the difference in perspective between medical researchers and evolutionary biologists. By perspective, I mean the types of explanatory frameworks one uses to understand a set of results.

Notice that I could have used the word "theory", but it actually has more to do with the cultural premises of one's discipline and formal training [1], especially in cases where there is a lack of good theory.

These slides are the second part of a talk of mine called "If your results are unpredictable, does it make them any less true? (posted to Figshare), which is a follow-up on the HTDE 2012 Workshop.

This set of slides was inspired by an in-lab discussion about a news article, that lead me to a recent PNAS paper on sepsis research in mice and humans. While mice are the accepted model organism for studying sepsis [2], it turns out that the physiological response (e.g. microarray studies and gene expression correlations) to sepsis in humans is very different than in mice. 

This result is interesting from an evolutionary standpoint. While there is phylogenetic distance between mice and humans, they are both mammals and certainly share many physiological and genomic characteristics. Furthermore, can these differences be explained using evolutionary theory? Has there been evolution in the sepsis response between mice and humans, or are these differences due to a highly variable response that can vary widely between species (and perhaps even between individuals in the same species)?

The variation in pathway activation and physiological responses seems to be quite common in medical research. When a certain experimental manipulation is done to multiple species [3], there is a range of possible outcomes, from a common response to a widely varying responses. We will return to this later. 

For now, let's consider why such massive differences might exist between humans and mice for a single physiological response. This is where we must return to the issue of premises. Given your background and preferences, you might choose a single explanatory framework. 

I have presented three in the slide below:  black box, complexity, and noise. Each of these may find support depending on the measures used and physiological components assayed. Yet each of them used in isolation may not be particularly satisfying, nor even explain very much of the data by themselves. This is why good, unified theories are of such value.

Another important aspect of understanding this variable response is to rule out alternative hypotheses. In the slide below, I consider three potential artifacts that could unduly influence the animal model results: standardization of environmental conditions, artificial selection on the model organism population due to selective breeding, and the tendency of the experimenter to put more weight on features of the experimental design or analysis that allow for greater experimental replication within a particular species. Particularly in the case of the first and last point, the lesson is that standardization of the experimental setting may actually do more harm than good and introduce ecological validity problems.

Now I present my interpretation of what is going on with the sepsis result. This consists of two hypotheses that can be applied to each species (human and mouse). The first is that the physiological response to sepsis is exact, which utilizes the same pathways and same patterns of gene expression across most conspecifics but only within a single species. This might require mutational distance and other evolutionary changes among the genes that explain the sepsis phenotype. 

The alternate hypothesis says that the physiological response to sepsis is variational, which means that there is potentially great variation is mechanism across most members of the same species. This variation need not be due to heritable mutation, but simply a lack of specificity in the molecular pathways and other associated mechanisms. In this case, there would be differences between human and mouse far greater than a consensus phylogeny might suggest.

What is a variational response? The term "variational" [4] is taken (perhaps loosely) from the mathematics and physics literature, and is generally used to describe a system with many potential solutions. In this context, the goal of the variational method is to approximate potential solutions based on optimizing their properties. 

One example can be shown in the slide below: two alternate routes from Toronto to Vancouver. Each route is the "shortest" route using two pathway criterion. One pathway is tightly restricted to the Trans-Canadian highway, while the other allows for an alternate route along a number of US interstates (e.g. 5, 90, 94). Both routes are about the same number of kilometers in length (e.g. number of steps in a physiological pathway). Yet they might be alternately used due to the in-capacitation of one pathway or the other [5].

The slide below shows these hypotheses in a phylogenetic context. As a contingency table, we consider the exact and variational scenarios for both conserved and divergent mechanisms. In the case of a conserved mechanism, there is very little mutational change to the underlying genes or pathway. For a divergent mechanism, the opposite is true.

To further understand what is meant by evolutionary conservation (and how it affects the consistency of physiological responses across species), I will now discuss two examples from the literature: the regulation of stress and aging, and the use of zebrafish as a human analogue. This will hopefully put my evolutionary speculations in context.

In aging research, phylogenetically-divergent species such as yeast and flatworms are used to understand the substrate of interventions such as caloric restriction and the activity of pathways related to stress resistance. As Longo and Fabrizio [6] demonstrate using aggregated data (see below slide), the associated pathway architectures are quite invariant across yeast, flatworms, and humans. However, this may not involve the same genes form species to species. In cases where conserved genes are known to be involved, it is not clear whether this conservation of mechanism components extends to a conserved mechanism itself.

A recent set of papers [7] focuses on comparing the genomes and proteomes of zebrafish with humans. As zebrafish and humans diverged around 440 million years ago [8], we would expect there to be vast differences in both function and genomics. However, there are occasionally greater differences among zebrafish than between zebrafish and humans. Another puzzle similar to the sepsis story, excpet that now we have extensive characterization of the genome and proteome to work from.

In the slide below (taken from Figure 3 of the Howe et.al paper), we can see how orthologues are shared by zebrafish and human as well as the relationship between so-called ohnologues in the zebrafish genome. Data such as these may provide good future estimates on how and why differences exist when evaluating variation related to basic physiological functions with and between zebrafish and humans.

So what can be learn from the big picture? Particularly when distinguishing between the homogeneity expected from experimental replication and the heterogeneity posed by natural variation? Perhaps we can treat experimental replication as a generative model, where the basic experiment is expected to reveal a range of likely outcomes. Like generative models in machine learning, the goal of analysis is to pick the best model (or in this case, the set of data that provide the closest match to what we know about the underlying natural phenomenon). 

This is a tricky proposition, because both the possible set of experimental and natural outcomes are incompletely known. Nevertheless, as in the case of understanding physiological processes and outcomes as variational processes, we can make good approximations that provide high explanatory power [9] without over-relying on the replication of results.

NOTES: 

[1] cultural premises are also known as "point of view". See this Tumbld Thoughts post for a detailed review.

From two wildly different premises: a painting entitled "Picasso and Dali Paint an Egg" (Artist unknown).

[2] model organisms are used conduct experiments that are either unethical or impossible to engage in with human subjects. Here is the full NIH list of model organisms. The accepted human analogues range from fruit flies (Drosophila) to mice (Mus musculus) and round worms (C. elegans) and zebrafish (Danio rerio). A newer trend is to use domesticated animals (e.g. sheep, cows, goats, pigs) as (non-traditional) model organisms.

Please see the following papers for more information on cross-species comparisons of model organisms with relevance to disease:

Golstein, P., Aubry, L., and Levraud, J.P.   Cell-death alternative model organisms: why and which? Nature Reviews Molecular and Cell Biology, 4(10), 798-807 (2003).

Goldstein, P.   Cell death in unusual but informative and beautiful model organisms. Seminars in Cancer Biology, 17(2), 91–93 (2007).

[3] this effect can be observed (usually understood via anecdotal reporting methods) both in vivo and in vitro (cell culture models). 

[4] the variational principle is widely used in quantum physics and engineering to arrive at solutions in very large, complex systems. Why not a version of this idea for physiological systems analysis?

[5] this suggests a role for mechanisms such as robustness, evolvability, and degeneracy.

[6] Longo, V.D. and Fabrizio, P.   Regulation of longevity and stress resistance: a molecular strategy conserved from yeast to humans? CMLS: Cellular and molecular life sciences, 59(6), 903-908 (2002).

[7] Here are a host of relevant papers (including a recent feature article in Nature):

a. Varshney, G.K. et.al   A large-scale zebrafish gene knockout resource for the genome-wide study of gene function. Genome Research, 23, 727-735 (2013).

b. Kettleborough, R.N.W. et.al   A systematic genome-wide analysis of zebrafish protein-coding gene function. Nature, doi:10.1038/nature11992 (2013).

c. Schier, A.F.   Zebrafish earns its stripes. Nature, doi:10.1038/nature12094 (2013).

d. Howe, K.   The zebrafish reference genome sequence and its relationship to the human genome. Nature, doi:10.1038/nature12111 (2013).

e. Barbazuk, W.B.   The Syntenic Relationship of the Zebrafish and Human Genomes. Genome Research, 10, 1351-1358 (2000).

[8] data derived from multiple consensus phylogenies (a meta-meta analysis) curated by Timetree.org. 

[9] This was cross-posted to my micro-blog, Tumbld Thoughts:

A new paper by Button et.al [a] featured in Wired Science claims that research in the Neurosciences are plagued with low statistical power (e.g. explanatory capacity of significant results), which is based on an 2005 paper by John Ioannidis [b] that applies a measure called positive predictive value (PPV) for determining the reliability of results in a particular scientific field (top image). While Ioannidis originally focused on results in Psychology, in later papers he has extended this line of inquiry to Computational Biology (e.g. microarray analysis) [c].

This reliability can be compromised by something called the proteus phenomenon [d], which deals with drawing a consensus from a series of datasets that exhibit similar biases. Two potential examples of this can be seen in a meta-meta analysis of the Psychological literature (Figure 3) from [a], and the Social Psychology literature. In the case of the latter, a paper from Vul et.al [e] investigates the exceedingly high correlations between brain activity data yielded from neuroimaging data and personality (e.g. self-reported) measures. Does this mean that there truly IS a high correlation, or is a subtle bias at work here?

Whether or not these concerns are overblown is up for debate, and it may be an artifact of the way we test for significance (e.g. NHST) rather than inherent problems with the method of experimental replication [f]. Fortunately, people are trying to address some of these issues (bottom image). Examples include the Equator Network [g] and the reproducibility project [h], both of which advocate open science. And, of course, there are more philosophically-oriented issues that I have started to address with the Hard-to Define Events (HTDE) approach.

For Reference 9, see also:

[a] Button, K.S., Ioannidis, J.P.A., Mokrysz, C., Nosek, B.A., Flint, J., Robinson, E.S.J., and Munafo, M.R.   Power failure: why small sample size undermines the reliability of neuroscience. Nature Reviews Neuroscience, doi:10.1038/nrn3475 (2013).

[b] Ioannidis, J. P.   Why most published research findings are false. PLoS Medicine, 2, e124 (2005).

[c] Ioannidis, J. P. et.al   Repeatability of published microarray gene expression analyses. Nature Genetics, 41, 149–155 (2009).

[d] Pfeiffer, T., Bertram, L. & Ioannidis, J. P. Quantifying selective reporting and the Proteus phenomenon for multiple datasets with similar bias. PLoS ONE 6, e18362 (2011).

"The chances for non-significant studies going in the same direction as the initial result are estimated to be lower than the chances for non-significant studies opposing the initial result"

[e] Vul, E., Harris, C., Winkielman, P.,  and Pashler, H.   Puzzlingly High Correlations in fMRI Studies of Emotion, Personality, and Social Cognition. Perspectives on Psychological Science, 4, 274 (2009). Also see Ed Vul's site on "voodoo correlations".

[f] for more information on the BEST test (an alternative to tests of the null hypothesis), please see: Kruschke, J.K.   Bayesian estimation supersedes the t test. Journal of Experimental Psychology: (2012).

[g] Simera, I. et.al   Transparent and accurate reporting increases reliability, utility, and impact of your research: reporting guidelines and the EQUATOR Network. BMC Medicine 8, 24 (2010). 

Other (more specialized) consortia geared toward this end include the Consolidated Standards of Reporting Trials (CONSORT), the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA), and the Collaborative Approach to Meta-Analysis and Review of Animal Data in Experimental Stroke (CAMARADES).

[h] Open-Science-Collaboration. An open, large-scale, collaborative effort to estimate the reproducibility of psychological science. Perspectives in Psychological Science, 7, 657–660 (2012).