The simulation approach in synthetic biology

https://doi.org/10.1016/j.shpsc.2013.03.010Get rights and content

Highlights

  • Synthetic biology aims at the de-novo design of generic objects like switches, oscillators, gates.

  • Rational design methods based on computational techniques are used to improve the design of generic objects.

  • Thus, Mathematical and computational strategies have been introduced to biology to tackle the problem of complexity.

  • The core question is: Will synthetic biology will merge with system biology due to the use of computational techniques ?

Abstract

Synthetic biology and systems biology are often highlighted as antagonistic strategies for dealing with the overwhelming complexity of biology (engineering versus understanding; tinkering in the lab versus modelling in the computer). However, a closer view of contemporary engineering methods (inextricably interwoven with mathematical modelling and simulation) and of the situation in biology (inextricably confronted with the intrinsic complexity of biomolecular environments) demonstrates that tinkering in the lab is increasingly supported by rational design methods. In other words: Synthetic biology and systems biology are merged by the use of computational techniques. These computational techniques are needed because the intrinsic complexity of biomolecular environments (stochasticity, non-linearities, system-level organization, evolution, independence, etc.) require advanced concepts of bio bricks and devices. A philosophical investigation of the history and nature of bio parts and devices reveals that these objects are imitating generic objects of engineering (switches, gates, oscillators, sensors, etc.), but the well-known design principles of generic objects are not sufficient for complex environments like cells. Therefore, the rational design methods have to be used to create more advanced generic objects, which are not only generic in their use, but also adaptive in their behavior. Case studies will show how simulation-based rational design methods are used to identify adequate parameters for synthesized designs (stability analyses), to improve lab experiments by ‘looking through noise’ (estimation of hidden variables and parameters), and to replace laborious and time-consuming post hoc tweaking in the lab by in-silico guidance (in-silico variation of bio brick properties). The overall aim of these developments, as will be argued in the discussion, is to achieve adaptive-generic instrumentality for bio parts and devices and thus increasingly merging systems and synthetic biology.

Introduction

For decades biology has tried to develop a fundamental understanding of biological processes in cells and organisms. It has collected exabytes of experimental data on biomolecular mechanisms and structures, but the complexity is so overwhelming that biology seems lost in floods of singular knowledge. Since 2000 two main strategies in dealing with the overwhelming complexity can be observed. On the one hand, data-based mathematical modelling and simulation are increasingly used to integrate the enormous amounts of experimental (omics) data and to create insights into the complex interactions of gene products, proteins, and cells (Wolkenhauer & Klingmüller, 2004). On the other hand, engineering principles have been proposed to simplify the ‘biological problem’ by breaking it down into controllable entities (Endy, 2005). Both strategies have become known as systems biology and synthetic biology. Synthetic biology is characterized as an ‘increasingly high-profile area of research […] towards the synthesis of living systems’, driven by the challenge of DNA-based device construction, genome-driven cell engineering and protocell creation (O’Malley, Powell, Davies, & Calvert, 2008, p. 57). It aims to engineer or design de-novo biological systems in the lab, which are useful for human purposes—since the ‘designs of natural biological systems,’ as a synthetic biology proponent alleged, ‘are not optimized by evolution for the purposes of human understanding and engineering’ (Endy, 2005, p. 450). Instead, systems biology attempts to understand biological systems using ‘mathematical concepts […] to illuminate the principles underlying biology at a genetic, molecular, cellular and even organismal level’ (Surridge, 2002, p. 205).

Both strategies are obviously antagonistic in their epistemic culture (science versus engineering), media (computer versus lab) and aims (understanding and discovery versus design)—requiring the ‘thing knowledge’ approach for synthetic biology (Gelfert, 2013, referring to Baird’s (2004) concept of thing knowledge) versus the ‘process knowledge’ for systems biology (Gramelsberger, 2013). Thing knowledge can be genuinely linked to the concept of controlled biological objects, called bio bricks. Bio bricks were introduced by the Massachusetts Institute of Technology, which established the Registry of Standard Biological Parts and the Synthetic Biology 1.0 conference in 2003, followed by the first International Genetically Engineered Machines (iGem) competition in 2004. The basic idea thereby is to divide the complicated problem of biological systems into simpler ones (decoupling) and then to rebuild ‘systems’ from them (designing). Literally spoken, one starts with ‘DNA bricks’ to build ‘parts’ and uses these ‘parts’ to assemble ‘devices’, and then takes the ‘devices’ to create useful biological ‘systems’ like bacteria producing drugs or other bio materials (Endy, 2005). In contrast to this thing/brick knowledge approach, the process knowledge approach of systems biology is rooted in the traditional scope of science, aiming at understanding based on empirical data and theoretical interpretation. ‘For this vision to succeed, we require foremost experiments and technologies that generate quantitative, time-resolved data’ (Wolkenhauer & Klingmüller, 2004, p. 22). These quantitative, time-resolved data are needed to investigate the dynamics of processes to rationally predict the complex behaviour of cells and organisms in order to discover purposeful biological systems.

The underlying question this paper poses is: Are synthetic biology and systems biology antagonistic strategies, or are they merging? In order to find an answer, the following line of argument will be explored: Both strategies are antagonistic if engineering in synthetic biology is understood as ‘tinkering’ or ‘kludging’ (O’Malley, 2011) in the lab. But can engineering really be understood as a purely lab-craft approach opposed to in-silico, theory-based methods? As well-established engineering domains such as vehicle construction show, they are based not on tinkering, but on ‘rational design’ (theory-guided, computer-based design methods). Today’s engineering is inextricably interwoven with mathematical modelling and simulation for design as well as for computer-aided manufacturing. There is hardly any part of a current airplane or car that does not emanate from the computer and from robotic manufacturing; not to mention the entire car or plane itself. However, if this advanced concept of engineering is linked to the ‘engineering of biology’ approach (CSynBI, 2009), the antagonistic setting disappears and areas of overlap appear between synthetic biology and systems biology. These overlapping areas are located in the use of computational techniques. In fact, ever since the very beginning in 2000 synthetic biologists have been using simulation techniques to optimize their de-novo designs. Vice versa, systems biologists try to create living microorganisms based on their in-silico concepts. Cases will be provided for both kinds of approaches in the following sections.1

Furthermore, the nature of controlled biological objects will be explored in greater depth. The ‘bricks’, ‘things’ or ‘objects’ of synthetic biology are not static, but dynamic. They have to function in living environments governed by noise, change and complexity. They are ‘living things’ themselves, which are forced to behave as well-defined machine parts. This creates an incongruity between living and technical objects, which has to be resolved by biology. One possibility of resolving this incongruity can be found, again, in engineering, because the current goal of engineering is to create new technical objects. Engineering and technology, in general, try to develop ‘generic objects’, which are general in their use and distribution and thus provide solutions for many problems (non-special-purpose objects as outlined in Shinn’s (2008) concept of research-technologies). Furthermore, new developments in engineering and technology aim at developing ‘adaptive objects’, which are open to self-modification in order to extend their intrinsic functional flexibility, and thus able to adapt to specific environments and uses—therefore implementing even more genericity. It is this kind of ‘adaptive-generic objects’ that are of interest for biology.

Following this line of thought, part two of this paper will outline the concept of generic and adaptive-generic objects in engineering. Based on cases, it will advance in part three to the need for simulation in synthetic biology and in part four towards a quantitative and predictive framework of engineering biology as a result of the convergence of the two approaches—systems and synthetic biology. In part five the development towards adaptive-generic instrumentality of bio parts will be discussed.

Section snippets

‘Generic’ and ‘adaptive-generic’ objects

The concept of genericity results from Terry Shinn’s characterization of ‘research-technologies’ as ‘base-line apparatus which can subsequently be transformed by engineers into products tailored to specific economic ends or by experimentators to further cognitive ends in academic research’ (Shinn, 2008, p. 10). Examples of this kind of precision instruments are the ultracentrifuge, the laser, the scanning tunneling microscope, but also conceptual tools like the Fourier transformation and the

Stability analysis

Interestingly, even the repressilator—the synthetic biology object of the first hour—could not have been created through pure tinkering in the lab. Because before any lab experiments could be done an in-silico stability analysis was necessary to identify areas in the parameter space of a genetic circuit that probably exhibit the desired behaviour and phenomenology. Tinkering would not be an effective strategy here, as it would cost weeks and months of work in the lab. Therefore stability

Towards a quantitative and predictive framework for engineering in biology

The cases demonstrate that systems biology as well as synthetic biology are trapped between the lack of quantitative experimental data for calibrating in-silico models and the lack of known variables and parameters for guiding lab experiments efficiently, which, in turn, requires the assistance of in-silico models. The link between the two is the lack of quantitative data, because bio parts and devices are not static objects but dynamic ones exhibiting specific process functionalities like

Discussion of adaptive-generic instrumentality for bio bricks

How can the diversity of molecular biological entities and functions be tackled? This core question drives systems biology as well as synthetic biology. As it has showed, synthetic biology picks up the idea of engineering desired functionalities, for which design principles are already known (based on negative and positive feedback mechanisms). It implements centuries-old concepts of generic objects like switches, gates, oscillators, sensors, and others into biomolecular networks that allow

Acknowledgements

The research was funded by the German Federal Ministry of Research and Education (01UB0925A). I like to thank the systems biologist and modeler Wolfram Liebermeister for his useful comments.

References (50)

  • G. Church

    From systems biology to synthetic biology

    Molecular Systems Biology

    (2005)
  • CSynBI (2009). Website of the centre for synthetic biology and innovation, Imperial College London....
  • H. de Jong

    Modeling and simulation of genetic regulatory systems: A literature review

    Journal of Computational Biology

    (2002)
  • A. de Silva et al.

    Molecular-scale logic gates

    Chemistry A European Journal

    (2004)
  • T. Ellis et al.

    Diversity-based, model-guided construction of synthetic gene networks with predicted functions

    Nature Biotechnology

    (2009)
  • M. Elowitz et al.

    A synthetic oscillatory network of transcriptional regulators

    Nature

    (2000)
  • D. Endy

    Foundations for engineering biology

    Nature

    (2005)
  • P. Galison

    Image and logic: A material culture of microphysics

    (1997)
  • T. Gardner et al.

    Construction of a genetic toggle switch in Escherichia coli

    Nature

    (2000)
  • G. Gramelsberger

    Generizität von Forschungstechnologien—mathematische und algorithmische Bedingungen

  • Gramelsberger, G. (2013). Simulation and systems understanding. In H. Andersen, D. Dieks, W. J. Gonzalez, T. Uebel, &...
  • A. Goldbeter

    Mechanism for oscillatory synthesis of cyclic AMP in Dictyostelium discoideum

    Nature

    (1975)
  • A. Hill et al.

    SynBioSS: The synthetic biology modeling suite

    Bioinformatics

    (2008)
  • M. Hucka

    The systems biology markup language (SBML): A medium for representation and exchange of biochemical network models

    Bioinformatics

    (2003)
  • Interview (2012). Bioquant. Center for quantitative analysis of molecular and cellular biosystems, University of...
  • Cited by (0)

    View full text