I just recently completed a review presentation about different systems architectures (e.g. feedback, feedforward control) found in biological systems. One example I am using is the futile cycle, which can be found in a number of biochemical systems.
To set the stage for this discussion, I am going to use a rather silly example from science fiction and relate it to real biological systems. Ever wonder why the machines might have used humans as batteries in the movie "The Matrix"? Or how birds can generate body heat on demand? In both cases, there are no great advances in bionics at work, just a set of forward and backward kinetic reactions called a futile cycle that produce nothing but heat energy as a by-product.
Some might argue that this is functionally pointless. For example, more energy would be expended feeding and maintaining each human than would be harvested as energy. However, in true Mythbusters fashion, a closer examination of futile cycle function might demonstrate why this fictitious example may not be so far-fetched, and why futile cycles might explain a lot about emergent complexity in biological systems.
Futile cycles consist of three processes: a forward reaction, a conversion the resulting product back into the original enzyme, and production of nothing but heat energy as an output. You can see an example of a futile cycle in the figure below. Hence, futility. Yet futile cycles may provide more than meets the eye, particularly when you consider the emergence, evolution, and context of the systems they are embodied within. I will now provide five examples of how futile cycles operate in and provide an opportunity for emergent complexity in biological systems.
Schematic of a typical futile cycle (example is from a metabolic pathway). COURTESY: Figure 1 in .
1) Futile cycles may be responsible for a facultative adaptation known as "thermoregulation on demand". In birds, a futile cycle reaction occurs in brown adipose tissue that allows for an "on-demand" heat source . The creation of a thermoregulatory switch from what might otherwise be considered a wasteful interaction is a good example of a dissipative structure . Dissipative structure are under-appreciated natural phenomena which may be addressed in a future post. The end result is that complex structure can emerge from and be supported by large fluctuations of energy (such as river systems or the bloodstream).
Qian and Beard  also suggest that futile cycles allow for high grade chemical energy to be converted into low grade heat energy. While this may appear to be a purely consumptive reaction, it actually enables different types of work to be done. One way to better understand this involves the change in information content resulting from a transformative process. In Samoilov, Plyasunov, and Arkin , it is suggested that colored noise (or other random fluctuations) can drive and even amplify the complex dynamics of a futile cycle. The relationship between noise, information content, and energy transformation is an interesting idea which I may explore in future posts.
2) Path-dependence are particularly important in maintaining the function of futile cycles, and is an important aspect of biological function in a number of systems. The path-dependence of components resembles the knocking-down of dominoes (see first figure below). A good example of this at the anatomical scale is the neural pathway in humans that governs striatal/basal ganglia function. In this case, regulation is contingent upon a carefully maintained sequence of disinhibitions (e.g. selective overriding of the default inhibitory state), which are ultimately key in enabling higher-level cognitive behaviors (for basic relationship, see second figure below). Futile cycles, particularly those that are coupled, are thought to enable complex processes at higher scales of organization in a similar fashion.
TOP: Sequence of dependent events (e.g. stacked dominoes). BOTTOM: neural circuit subject to sequential regulation. COURTESY: Figure from .
If this sequence of events is disturbed, futile cycles (and likewise neurobehavioral regulation) may no longer be able to function. However, in many cases, futile cycles may not be specific enough to exhibit this property. If all you need is a set of enzymes to provide a beneficial reaction, then there might be a large class of enzymatic pathways from which to choose.
3) How might a futile cycle arise in evolution? Perhaps in the course of producing two products needed by the cell in at alternating points in time. Viewed from an ecological perspective, the enzymes involved in the futile cycle become interdependent. For example, a futile cycle in one system (species A) might become the substrate for another system (Species B). One niche evolves on top of an established population, which can be seen in many bacterial and viral populations.
The resulting relationship can either be parasitic (where species B benefits over species A), mutualistic (both species A and B benefit" from the relationship). This ecological relationship enabled by one or a series of futile cycles is based on the creation and maintenance of a dissipative structure, while the species (B) that always benefits also exhibits characteristics of niche construction. Niche construction  is particularly interesting in that it plays an integral role in regulating the dynamics of coevolutionary relationships.
4) Futile cycles, or rather their recursive nature, may serve as generalized homeostatic mechanisms. One example is the seesaw model (see figure below), in which the degradation of one product initiates the production of the second product and vice versa. When viewed over time, this results in a mutual pulsing [see 7 for an example form cell cycle]. In the case of MPF and cell cycle, where MPF is produced and depleted cyclically, the basic cycle motif is dependent upon the initiation of a process (e.g. phase in cell cycle).
Eukaryotic cell cycle functioning as a futile cycle.
5) Interconnected futile cycles -- in a larger-scale context, futile cycle may be interconnected. In these cases, the effects can be either processive or distributive. In isolation, futile cycles are processive, meaning that one input gives rise to a single set of reactions with one output. Upon linking together multiple futile cycles, however, a single input can give rise to a set of reactions with multiple potential outputs. Recall that this is related to the issue of path-dependence and functional stability. Wang and Sontag  have considered the behavior of interconnected futile cycles, and have proposed that when interconnected, such systems exhibit unique steady states. This can be particularly useful in enabling processes such as phosphorylation, dephosphorylation, and MAPK cascades, all of which are key in regulating biochemical signaling.
From early concepts of the second of law of thermodynamics to treatments of entropy in modern cosmological theory, the depletion of free energy in a complex system has been thought of as "heat death". And yet the connection between futile cycles and thermodynamic processes could be more accurately described as "heat life". While I would not make the claim that futile cycles break the second law of thermodynamics, they do provide a mechanism and substrate for building up complexity in the face of entropy.
The purpose of this post is to facilitate the cross-pollination of ideas and design principles between hardware customizers and synthetic biologists, in addition to exploring the nature of open-source innovation. What can open-source hardware do to inform the design of synthetic biological systems, and vice versa? And what are the latest trends in do-it-yourself science and technology?
A movement is afoot to make custom electronics hardware. While this has been a hobby almost as long as there have been electronics, the focus has always been on hacking proprietary designs. Now the open-source hardware movement is an attempt to build hardware according to the same principles governing open-source software. Open-source software is of course where users all contribute their own modification designs to the general community, with a systems design that allows for this to be done at low cost. The official working definition of open-source hardware is "hardware whose design is made publicly available so that anyone can study, modify, distribute, make, and sell the design or hardware based on that design".
The Maker Faire in New York is having a spinoff conference called the Open Hardware Summit. To encourage interest in this area, the Makezine Blog has provided a few examples of open hardware designs. There seem to be two versions of the concept. One is an open model through which design are provided online (e.g. the motor-driven screw block example) and people can assemble the actual device or its variants on their own. The other is API-driven, such as the Roomba, which allows people to customize a basic hardware design using custom components and software libraries.
There is a parallel open-source movement in biology. While most of the work to date has focused on microbial bioengineering, many of the lessons learned have broader application. One goal of this movement is to establish a catalog of standard biological parts (called " biobricks"). This catalog is organized by types of parts, whichinclude DNA, translational units, and terminators.
A parts registry also has broader application in the systems biology community. For example, the engineering of animal cells requires characterization and analysis of transcriptional networks and signaling pathways. Yet even with a good parts catalog, open-source modification of biological systems can be hard. The tinkerer is essentially taking something that has evolved through natural processes and makes customized modifications with relatively little control over the outcome.
Finally, let us recall the working definition of open-source hardware: "hardware whose design is made publicly available so that anyone can study, modify, distribute, make, and sell the design or hardware based on that design". There are two comparisons that can be made between hacking inorganic and organic systems:
1) biological modification must take into account natural diversity between the individual unit being modified. In an inorganic system (e.g. a toaster), the hardware template is most likely to be identical from one unit to another. In biological systems, there is much (undocumented) diversity between units. Some of this is structural (e.g. DNA or protein mutations), but much of it is also functional (e.g. biological rhythms or stochastic gene expression).
2) the catalog of standard parts approach might not fully enable innovation in either organic or inorganic systems. This is an interesting thought that I have not fully explored, but may be due to parallels between human innovation and the complex nature of biological systems. For inorganic systems, the innovation of modification often exceeds what is written in the instruction manual. This process of cultural adaptation may have strong parallels with the adaptable substrate of a biological system.
I am interested in people's thoughts on this topic, particularly if there are any more parallels between organic and inorganic systems that I have not discussed here.
As an academic with a complicated and discontinuous academic heritage, I've often wondered how communication (internet) technologies and the nature of increasing academic interactions across disciplinary boundaries will affect the evolution of different disciplines. Typically, the transmission of academic information and ideas has passed from mentor to protégé in a way that has lead to the formation of distinct departments and disciplines. Current and future trends are likely to obscure this relationship, at least with regard to how knowledge and points of view are acquired.
The mentor-protégé model is an idealized view of how academia operates in practice. Yet it could be argued that the relative obscurity/advanced nature of academic knowledge have made alternate systems of handing down knowledge and certifying expertise hard to implement and maintain. While communication (internet) technologies can reduce these burdens and facilitate interactions between academics, is it enough to fundamentally change the academic enterprise?
The hierarchical nature of the mentor-protégé relationship is captured by an academic genealogy [link]. For those who are unaccustomed to an academic genealogy, think of a family tree that is based on whom a person has studied under and their academic mentors. A succession of mentors can go back a number of generations, even a number of centuries. One example is from the Mathematics Genealogy Project, which I provide for a sense of scale and topology
A second example is from NeuroTree (with help from FlyTree and ChemistryTree) centered upon Thomas Morgan Hunt, a well-known fly geneticist. The example below show roots going back to Georges Cuvier (a famous natural historian from the late 18th and early 19th century), and a massive proliferation of leaves extending to our time. Thomas Hunt Morgan trained many scientists because he was prolific, while his work influenced many more academics not shown in this tree.
What I would like people to notice is the relatively strict hierarchy in these trees. By and large, the flow of expertise goes in a single direction (from mentor to protégé), and often does not reticulate (rejoining disparate parts of the tree). This was before the internet age, which has affected the retrieval of information and access to publications profoundly. We can see from the figure below that the topology of the internet is highly decentralized. While there is some hierarchical structure, most if not all nodes take in multiple inputs. The topology and scale of the internet network is shown in the image below.
However, the internet has yet to transform the basic model of training academic scientists. I suspect that this is changing, however, and in a few generations scientific training will be much different. As a rule, traditional disciplines are ever-changing entities. Cross-disciplinary training programs arise when there is a need for training based on different perspectives. I suspect that even today, the genealogy model of academic influence is highly contrived. People have been reading the broader scientific literature for centuries, and some disciplines/departments are more closed-off than others. Thus, the degree of strict hierarchy is determined in part by academic context.
In the case of NeuroTree, you can observe occasional reticulations in the tree topology. However, using the criterion of mentor-protege as the sole criterion for creating a link limits the number of observed interactions. The Erdos Number Project is an example of academic influence measured not by academic supervisor, but on article co-authorship. Viewing relationships between generations of academics in this way yields a bi-directional, cyclic graph structure more capable of capturing interactions. In the picture below, we can see that your “degree”, or distance in links from the mathematician Paul Erdos is related to whether you or your colleagues have published a paper with Erdos. So if a co-author of a paper on your publications list also worked on a paper with Erdos, your Erdos number is 2. These networks ultimately exhibit a small-world effect, where all members of the community are only several links away from everyone else. This does not preclude a hierarchical relationship between mentor and protégé. It only uses a different means of measuring influence.
In conclusion, academic genealogies are most likely an incomplete way of determining influence or even expertise on the part of a given academic. This is partially because knowledge is acquired from many sources, not just a mentor. On the other hand, a mentor can influence their trainees in much the same way a parent culturally influences their offspring. In these cases, a mentor’s influence can go hand in hand with the surrounding culture of the local department/University in reinforcing certain schools of thought. Which make this topic a fascinating case study in cultural evolution.