Section 7: Networks
The complex biological molecules we have spent the previous sections trying to understand are the building blocks of life, but it is far from obvious to put these building blocks together into a coherent whole. Biological molecules, as well as cells and complete organisms, are organized in complex networks. A network is defined as a system in which information flows into nodes, is processed, and then flows back out. The network's output is a function of both the inputs and a series of edges that are the bidirectional paths of information flow between the nodes. The theory and practice of networks is a vast subject, and with one ultimate goal of understanding that greatest of mysteries, the human brain. We will return to the brain and its neural networks in the next section. For now, we will discuss the more prosaic networks in living organisms, which are still complex enough to be very intimidating.
It isn't obvious when you look at a cell that a network exists there. The cytoplasm of a living cell is a very dynamic entity, but at least at first glance seems to basically be a bag of biological molecules mixed chaotically together. It is somewhat of a shock to realize that this bag of molecules actually contains a huge number of highly specific biological networks all operating under tight control. For example, when an epithelial cell moving across a substrate, patterns of specific molecules drive the motion of the cell's internal skeleton. When these molecules are tagged with a protein that glows red, displaying the collective molecular motion under a microscope, a very complex and interactive set of networks appears.
Figure 26: The circuit diagram (top), bacterial population (center), and plot of the dynamics (bottom) of the repressilator, an example of a simple synthetic biological network.
Source: © Top: Wikimedia Commons, Public Domain. Author: Timreid, 26 February 2007. Center and bottom: Reprinted by permission from Macmillan Publishers Ltd: Nature 403, 335-338 (20 January 2000). More info
The emergent network of the cytoplasm is a system of interconnected units. Each unit has at least an input and an output, and some sort of a control input which can modulate the relationship between the input and the output. Networks can be analog, which means that in principle the inputs and outputs are continuous functions of some variable; or they can be digital, which means that they have finite values, typically 1 or 0 for a binary system. The computer on which this text was typed is a digital network consisting of binary logic gates, while the person who typed the text is an analog network.
There are many different kinds of biological networks, and they cover a huge range of length scales, from the submicron (a micron is a millionth of a meter) to the scales spanning the Earth. Outside of neural networks, the most important ones are (roughly in order of increasing abstraction):
- Metabolite networks: These networks control how a cell turns food (in the form of sugar) into energy that it can use to function. Enzymes (proteins) are the nodes, and the smaller molecules that represent the flow of chemical energy in the cell are the edges.
- Signal transduction networks: These networks transfer information from outside the cell into its interior, and translate that information into a set of instructions for activity within the cell. Proteins are the nodes, typically proteins called kineases, and diffusible small signaling molecules which have been chemically modified are the edges. This is a huge class of networks, ranging from networks that process sensory stimuli to chemical inputs such as hormones.
- Transcriptional regulatory networks: These networks determine how genes are turned on and off (or modulated).
- Interorganism networks: This is a very broad term that encompasses everything from the coordinated behavior of a group of bacteria to complex ecologies. The nodes are individual cells, and the edges are the many different physical ways that cells can interact with each other.
There are probably fundamental network design principles that must be obeyed independent of their biological or manmade (which is still biological) origin if the network is to be stable to perturbations. Instability in a network is not generally viewed as a good thing, although there are exceptions to this rule. For example, the aerodynamics of most modern fighter jets makes the plane inherently unstable. This sounds like a very bad thing, except that it makes the fighter extremely adaptive to direction changes. Modern computers can constantly monitor and instantaneously correct the instability, so we end up with aircraft that are far more maneuverable—and thus more effective fighters—than the ordinary, stable variety.
The kind of stability issues the fighter jet and other similar networks face are deterministic, and can be modeled by ordinary differential equations that are straightforward to write down. One might then imagine designing a network based on a set of these equations. One of the pioneering exercises in designing "from scratch" was the work of Elowitz and Leibler of an oscillating gene expression pattern. It is sobering to understand the depth of understanding that was necessary to have made this simple oscillator work. For an electrical engineer, it is straightforward to design an oscillator following some basic rules of electromagnetism. However, in a biological network, the parameters are much less cleanly defined. Despite the inherent challenges, we now have a basic set of biological modules that is being developed in the new field of "synthetic biology," which is a mix of physics, engineering, and biology that exploits our knowledge of networks to design new functional biological "circuits." Figure 27 shows an example of a biological "film" consisting of bacteria. To do this, a gene was inserted into E. coli that coded for a protein that causes the bacteria to make a black pigment. The pigment production was coupled to a light sensor, so that pigment would be made only in the dark. The group used stencils to pattern light exposure and produce bacterial photography.
Figure 27: Portrait of the University of Texas at Austin (UT Austin)/University of California, San Francisco (UCSF) Synthetic Biology team. The left panel shows the projected image of the students and professors from UT Austin and UCSF who participated in the project, and the right panel shows the resulting bacterial photo.
Source: © Aaron Chevalier and the University of Texas at Austin. More info
In addition to deterministic stability issues in biological networks, there is also the issue of stability in the presence of noise. For example, at the macro-scale, the sensory network of the dog's nose is about a million times more sensitive than a human noise. Despite this extreme sensitivity, the dog nose is not overwhelmed by the presence of an enormous background of other molecules. That is, the dog nose sensory network is extremely good at noise rejection. At the micro-scale of the single cell, the very finite number of molecules actually involved in the network node edges leads to statistical noise that can either confound the network's stability or increase its sensitivity. The dog has clearly resolved this issue to its benefit, but it remains a problem in the design of synthetic biological networks.
The obvious end goal of synthetic biology could be something truly astonishing: synthetic life. The key to synthetic life, if it is indeed achieved, will be our ability to harness biological networks. So far, scientists have synthesized genetic material and other important biological molecules from scratch, but have not put the pieces together into a complete living organism. The feat hailed by the media as the so-called creation of a "new form of life" by Craig Venter is something of a misnomer. While Venter's synthesis of a functional bacterial chromosome of one million base pairs was a fantastic technical achievement, it is very far from synthetic life, as the new chromosome was inserted into an already functioning cell consisting of a enormous collection of interacting networks, which we neither can understand nor can reproduce. Until we can understand the emergence of life from the networks of biology, we will remain very far from achieving synthetic life.