Computation in Cells and Tissues

Cells are complex systems. For some people, they are like tiny chemical plants, or laboratories, or machines. For others they are more like computational devices or even computational systems. As data continue to be generated about them and their components and the systems that they make up, new perspectives and models are needed to deal with the complexity. Cells are more than bags of chemicals just as macromolecules are more than microscopic billiard balls or strings of beads. The challenges of an information processing view that complements the more commonly expressed chemical processing view needs to be taken into account.

There are many examples of the natural world providing inspiration for human engineers and designers. Cell biology is one branch of the natural sciences which has not yet been widely exploited in this way, but which has great potential for application, particularly in the telecommunications area. The features of cells map strikingly well on to some of the challenges in current engineering and design for telecommunications systems. The autonomy, evolution, adaptivity and self-organisation of cells are all desirable for the complex, dynamic and geographically distributed networks we are now constructing and using. Three examples of current research illustrate how analogies from cells can lead to radically different telecommunications systems. Cell fate behaviour in fruitfly cells has inspired a new, decentralised approach to managing mobile phone networks. Morphogenetic events in early embryos point to new design methods which add depth to the established, and also biologically inspired, techniques of evolutionary optimisation. Genetic control pathways in bacteria inspire implicit learning techniques which allow individual ‘cells’ in simulation to discover adaptive behaviour without an explicit definition of ‘fitness’. All of the examples are at the research stage, and will not be used in real networks until 2004 and beyond. However, they give a glimpse of the strength of the analogy between biological cells and elements in telecommunications systems, and suggest that this will be a productive area of work into the future.

The immune system is highly distributed, highly adaptive, self-organising in nature, maintains a memory of past encounters and has the ability to continually learn about new encounters. From a computational point of view, the immune system has much to offer by way of inspiration to computer scientists and engineers alike. As computational problems become more complex, increasingly, people are seeking out novel approaches to these problems, often turning to nature for inspiration. A great deal of attention is now being paid to the vertebrate immune system as a potential source of inspiration, where it is thought that different insights and alternative solutions can be gleaned, over and above other biologically inspired methods.

Given this rise in attention to the immune system, it seems appropriate to explore this area in some detail. This survey explores the salient features of the immune system that are inspiring computer scientists and engineers to build Artificial Immune Systems. An extensive survey of applications is presented, ranging from network security to optimisation and machine learning. However, this is not complete, as no survey ever is, but it is hoped this will go some way to illustrate the potential of this exciting and novel area of research.

This first part of this article details and expands the work on embryon-ics, a recently proposed fault-tolerant cellular architecture with reconfiguration properties inspired by the ontogenetic development of multicellular systems. The design of a selector-based embryonic cell and its applications are presented. The second part of this article describes a novel approach to hardware fault tolerance that takes inspiration from the human immune system as a method of fault detection. The human immune system is a remarkable system of interacting cells and organs that protects the body from invasion and maintain reliable operation even in the presence of invading bacteria or viruses. Here we seek to address the field of electronic hardware fault tolerance from an immunological perspective with the aim of showing how novel methods based upon the operation of the immune system can both complement and create new approaches to the development of reliable hardware systems. The final part of the article suggests a combined architecture that would have the characteristics and advantages of both Embryonics and immunotronics.

The functional integration of man-made devices and biological systems represents one of the grand challenges of science and technology. Efficient real-time exchange of information and/or materials across the molecular-scale interface between biological and physical systems is a core platform requirement to realise that vision. This common technology requirement underpins development of (i) affordable diagnostic devices that harness the full potential of genomic information through real-time predictive, preventive, point-of-care and personalised health care provision; (ii) anti-terrorism, environmental, food, crime detection and process monitoring sensors; (iii) targeted drug delivery systems; (iv) advanced orthopaedic and neural implants; (v) pharmaceutical screening and lab-on-chip devices; and (vi) ubiquitous systems which monitor, interact with, and respond to biological events, and link unobtrusively with information processing and communications systems. A thematic area common to this huge diversity of devices, applications and sectors, each of which on its own could form the subject of one or more integrated projects, provides the focus of the Network for Biomedical Applications of Micro and Nano Technologies (NANOMED), which will not only advance the associated platform science and technology, but act to link diverse communities.

‘Bioinformatics’ is used to describe computational topics in molecular and cellular biology. As a discipline it involves cross-fertilisation of ideas between computer science and modern biology. DNA, RNA and protein are classes of macromolecule whose members play several roles including inheritance, biological information processing, signal transduction and catalysis. Methods of classifying these molecules are central to current methods for elucidating relationships between sequence, structure and function. We take as a case study metaphors for the function of proteins and point to a unified view of proteins as computational devices capable of matching patterns as inputs and processing to result in alternative outputs. Finally we consider the requirement for a systems view of life in order to construct new models for the era of post-genomic biomedicine. The subject has an ethical dimension and we consider the case that such models are metaphoric constructions.

This chapter provides a short review of the modelling of Genetic Regulatory Networks (GRNs). GRNs have a basic requirement to model (at least) some parts of a biological system using some kind of logical formalism. They represent the set of all interactions among genes and their products for determining the temporal and spatial patterns of expression of a set of genes. The origin of modelling the regulation of gene expression goes back to the Nobel-prize winning work of Lwoff, Jacob and Monod on the mechanisms underlying the behaviour of bacterial viruses that switch between so-called lytic and lysogenic states. Some of the circuit-based approaches to GRNs such as the work of Kauffman, Thomas, and Shapiro and Adam are discussed.

Evolution has frequently been seen as a result of the continuous or discontinuous accumulation of small mutations. Over the many years since Darwin, it has been found that simple point mutations are not the only mechanism driving genomic change, for example, plasmids, transposons, bacteriophages, insertion sequences, deletion and duplication, and stress-sensitive mutation all have a part to play in directing the genetic composition and variation of organisms towards meeting the moving target that is the environmental ideal that exists at any one time. These generate the variation necessary to allow rapid evolutionary response to changing environmental conditions.

We review some of our recent results on establishing a neuronal decision theory and spiking ICA (independent component analysis). For neuronal decision theory, we show that the discrimination capacity of a model neuron is a decreasing function of inhibitory inputs. Increasing the output variability of neuron efferent firings implies an improvement of neuron discrimination capacity. For the two most interesting cases, with or without inhibitory inputs, the critical discrimination capacity is exactly given. For spiking ICA, by a simple combination of the Informax principle and the input-output relationship of a spiking neuron, we first develop a learning rule. By applying the learning rule to linear mixture of signals, we demonstrate that spiking neuron network can accomplish ICA tasks.

Pattern formation in multicellular organisms generally occurs within populations of cells that are in close contact. It is thus natural and important to consider models of pattern formation that are constructed using a spatially discrete cellular structure. Here, the particular case of pattern formation in cellular systems that depends on contact-dependent (juxtacrine) signalling between cells is discussed. Spatial and spatio-temporal patterns can emerge in populations of cells coupled by juxtacrine signalling when the degree of activation of the relevant cell-surface receptors regulates both the pathway of differentiation adopted by the cell and the ability of the cell to participate in further juxtacrine signalling. When this latter condition applies, juxtacrine signalling couples all the cells of a population to form a spatially extended signalling network. Due to the essential nonlinearity of the signalling, such juxtacrine networks can exhibit dynamics that are quite different to those in networks of cells coupled by linear diffusion. Two simple cases are discussed here, in which receptor activation either diminishes or enhances the signalling ability of a cell. In the former case, signalling can act to amplify small differences between cells via a feedback-mediated competition, leading to stable spatially periodic patterns (a process known as lateral inhibition). In the latter case, signalling can result in a range of different patterns, including stable spatial gradients, propagating fronts, and periodic and quasi-periodic spatial patterns. These quite simple examples serve to illustrate the potential richness of this important class of biological signalling, and provide guidance for the development of more complex models.

This chapter describes a model of the hypothalamic and pituitary components involved in controlling growth hormone release. The model has been developed by gathering and attempting to formalise the experimental data on the system but has been kept as simple as possible, focusing on the functional rather than mechanical properties of its components. In this way it has shown that a relatively simple model can be capable of producing complex behaviour and accurately reproducing the behaviour and output of a real brain system.

Computational models have been of interest in biology for many years and have represented a particular approach to trying to understand biological processes and phenomena from a systems point of view. One of the most natural and accessible computational models is the state machine. These come in a variety of types and possess a variety of properties. This Chapter discusses some useful ones and looks at how machines involving simpler machines can be used to build plausible models of dynamic, reactive and developing biological systems which exhibit hierarchical structures and behaviours.

Ciliate is a term applied to any member of a group of around 10,000 different types of single-celled organisms that are characterised by two unique features: the possession of hair-like cilia for movement, and the presence of two nuclei instead of the usual one. One nucleus (the micronucleus) is used for sexual exchange of DNA, and the other (the macronucleus) is responsible for cell growth and proliferation. Crucially, the micronucleus contains an “encoded” description of the working macronucleus, which is decoded during development. This encoding “scrambles” functional gene elements by both the permutation of coding sequences and the inclusion of non-coding sequences. A picture of the ciliate Oxytricha nova is shown in Fig. 1. During development, ciliates reorganise the material in the micronucleus by removing non-coding sequences and placing coding sequences in the correct order. This ‘unscrambling’ may be interpreted as a computational process during which up to 95% of the original sequence is discarded. The exact mechanism by which genes are unscrambled is not yet fully understood. We first describe experimental observations that have at least suggested possible mechanisms. We then describe two different models of the process. We conclude with a discussion of the computational and biological implications of this work.

This chapter considers a number of ways in which individual molecules in protein signalling systems can be thought of as computational agents. We begin with a general discussion of some of the ways proteins can be viewed from an information processing point of view. The degree of computational prowess shown by many proteins, such as enzymes and transcription factors is discussed; specifically in terms of a number of ‘cognitive’ capacities. We review some of the proteins involved in signalling that make use of transfer of phosphate groups (kinases and phosphatases) and focus attention on the Yeast MAP Kinase cascades. An algebraic approach to modelling certain aspects of protein interactions is introduced. We begin with a simple algebraic model which we describe in some depth, using yeast signalling pathways as an example; we then describe techniques and tools which promise more sophisticated models.

The aim is to explain and explore some of the current ideas from category theory that enable various mathematical descriptions of hierarchical structures. We review some aspects of the history and motivations behind the development of category theory and how it has impacted on developments in theoretical biology and theoretical computer science. This leads on to a discussion of hierarchical systems and a discussion of some simple examples. The important idea of colimit is then introduced. Towards the end of the chapter a number of open questions and problems are discussed.

Although it is common knowledge that cellular systems are dynamic and regulated processes, to this date they are not investigated and represented as such. The kinds of experimental techniques, which have been available in molecular biology, largely determined the material reductionism, which describes gene expression by means of molecular characterisation.

Instead of trying to identify genes as causal agents for some function, role, or change in phenotype we ought to relate these observations to sequences of events. In other words, in systems biology, instead of looking for a gene that is the reason, explanation or cause of some phenomenon we seek an explanation in the dynamics (sequences of events ordered by time) that led to it.

In mathematical systems biology we are aiming at developing a systems theory for the dynamics of a cell. In this text we first define the concept of complexity in the context of gene expression and regulation before we discuss the challenges and problems in developing mathematical models of cellular dynamics, and provide an example to illustrate systems biology, its challenges and perspectives of this emerging area of research.

This chapter is concerned with how computational ideas can be used as the basis for understanding biological systems, not by simulating such systems, but by taking a computational stance towards the way such systems work. A number of issues are addressed. Firstly the question of what kinds of computer science are needed to help understand computational processes which happen outside of conventional computing machines. The second issue addressed places computational constraints on how the world can act into Dennett’s framework of grades of possibility. The final main section considers the issue of changes in the world, and when it is meaningful to regard such changes as carrying out computations.