Analysis of Biological Systems

The following sections are included:

• Dedication
• Preface
• Acknowledgments
• Contents

The convergence between computer science and biology has occurred in successive waves, involving increasingly deeper concepts of computing. Since its early days, computer science has taken inspiration from nature, with the works of Turing, Von Neumann and Minsky. These milestones led to extraordinary results, some of which recall biology even in their names: cellular automata, neural networks, genetic algorithms. The current situation makes computer science a suitable candidate for becoming a foundation for systems biology with the same importance as mathematics, chemistry and physics. The same applies to systems biology domains related to the healthcare sector such as systems nutrition (which studies the impact of nutrients on cellular machinery) or pharmacology…

This chapter introduces the molecular components and processes of the cell, which form the systems that are modeled in systems biology. We also present a number of laboratory techniques that are commonly used to produce data in systems biology. The material presented here focuses on eukaryotic cells, and specifically on human systems wherever speciesspecific detail is provided. As this is intended to be an introduction to cell and molecular biology, this chapter does not delve deeply into these topics, and also does not strictly require background knowledge in biology. Additional references are provided at the end of the chapter to provide the reader with resources for further information on these topics.

This chapter introduces the notions of systems and models, and also discusses their main properties.

In this chapter we present static data analytical approaches for systems biology. The goal of this chapter is to provide the reader with a practical introduction to a variety of statistical and bioinformatic approaches to analysis of biological data. Some of the methods presented are specifically designed for high-throughput data; however, many of the approaches are broadly applicable to smaller datasets as well. To contain this material within a single chapter, we do not present a deep theoretical treatment of the presented methods. Moreover, we are not able to discuss every method in statistics and bioinformatics.We instead provide a practical introduction to linear regression, dimensionality reduction and pathway/network analysis, as we feel that the methods presented in these sections will allow the reader to approach analysis of a dataset from a number of angles, and will also provide the reader with a foundation for understanding more specialized or advanced techniques.

We survey the main technologies used for modeling the dynamics of biological systems. We start with the more frequently used equation-based approaches that rely on classical mathematics. Then we move to networkbased modeling that has a mixed background in mathematics and engineering. Rewriting systems and automata-based modeling are derived from formal language theory. Finally, we introduce graphic notations. Algorithmic modeling is dealt with in the next chapter. Hereafter, we will focus on molecular interaction modeling, if not otherwise specified. We will assume that the outcome of a simulation is the dynamic behavior of a model, i.e., to approximate reproduction of time behavior of a real system (see Chapter 8).

We introduce modeling technologies based on programming languages under the assumption that biological elements are represented by pieces of code (processes, programs) and their interaction is modeled by message passing/data sharing or scope redesign of the corresponding programs. Dynamic evolution of biological systems is then represented by the state changes of the execution of the programs. The very first approach dates back to Fontana and Buss, who used λ-calculus to model biological systems. Subsequent works based on process calculi had more dissemination. First we introduce process calculi and their derived languages, then we introduce domain-specific languages (their formal semantics is given in Appendix C).

The modeling process starts from the definition of the objective of the model and the criteria for the model's being accepted by the users. The acceptance criteria are fundamental because they set the termination condition for the modeling process…

Simulation is the process of model execution to reproduce the dynamics of a system. Digital simulation (hereafter just simulation because it is the kind we focus on) is the process of making a computer behave like a system of interest and therefore relates models and computers. Modeling is the process of creating a model and of connecting it to a real system, and simulation is the process of connecting a model to a digital computer.

Investigating the mechanistic properties of a system through simulation is faster (e.g., many cell cycles of mammalian cells can be simulated in seconds on a computer, while it takes days in a lab) and cheaper (e.g., simulating crash tests on a computer avoids destroying physical objects or simulating the effect of perturbation on biological systems may save animals).

The outcome of a simulation is the dynamic behavior of a model, i.e., the approximate reproduction of time behavior of a real system. If we model a disease and we find, through simulation, a drug that can cure the disease or even prevent it, we cannot be sure that the result would be the same in the real world. We need to follow the protocol for drug validation before delivering it to real patients. The advantage of simulation is discarding alternatives that will not work and helping experimenters to design focused experiments, thus saving time and money.

This chapter is organized as follows. The next section introduces the basic notions of model execution. Then, we discuss how to build random number generators that are the basis of each simulator.We then concentrate on discrete-event simulation and the stochastic algorithms that are the core of stochastic simulation engines.

This book has introduced techniques (both static and dynamic) to produce models of biological systems with the aim of understanding, predicting and controlling their behavior…

The following sections are included:

• Appendixes
• Bibliography
• Index