A First Course in Systems Biology. ISBN 0815345682, 978-0815345688
A First Course in Systems Biology. ISBN 0815345682, 978-0815345688
A First Course in Systems Biology. ISBN 0815345682, 978-0815345688
Visit the link below to download the full version of this book:
https://cheaptodownload.com/product/a-first-course-in-systems-biology-2nd-editio
n-full-pdf-download/
To Ann,
Still the Hub of my Support System
A F IR S T COUR SE IN
SYSTEMS
BIOLOGY
SECOND
EDITION
Eberhard O. Voit
Garland Science Front cover image. The beautiful geometric shape of the fractal is
Vice President: Denise Schanck called self-similar because it has the same appearance at smaller
Senior Editorial Assistant: Katie Laurentiev and smaller scales. It reminds us of fundamental design features like
Assistant Editor: David Borrowdale feedback loops that we encounter at many organizational levels of
Production Editor: Georgina Lucas biological systems. Fractals are generated with nonlinear recursive
Illustrations: Nigel Orme models, and they are discussed with simpler examples in Chapter 4.
Copyeditor: H.M. (Mac) Clarke (Courtesy of Wolfgang Beyer under Creative Commons Attribution-
Typesetting: NovaTechset Pvt Ltd Share Alike 3.0 Unported license.)
Proofreader: Sally Huish
Indexer: Nancy Newman
Cover Design: Andrew Magee
ISBN 978-0-8153-4568-8
Published by Garland Science, Taylor & Francis Group, LLC, an informa business,
711 Third Avenue, 8th Floor, New York NY 10017, USA,
and 2 Park Square, Milton Park, Abingdon, OX14 4RN, UK.
15 14 13 12 11 10 9 8 7 6 5 4 3 2 1
Hard to believe, but it is already time for the second edition! I am happy to report
that the first edition of A First Course in Systems Biology has met with great suc-
cess. The book has been a required or recommended text for over 70 courses
worldwide, and it has even been translated into Korean. So why should a new
edition be necessary after only five short years? Well, much has happened.
Systems biology has come out of the shadows with gusto. Research is flourishing
worldwide, quite a few new journals have been launched, and many institutions
now offer courses in the field.
While the landscape of systems biology has evolved rapidly, the fundamental
topics covered by the first edition are as important as they were five years ago
and probably will be several decades from now. Thus, I decided to retain the
structure of the first edition but have rearranged some items and added a few
topics, along with new examples. At Georgia Tech we have used the book to
teach well over 1000 students, mostly at the undergraduate level, but also for an
introductory graduate-level course. Most of the additions and amendments to
this new edition respond to feedback from these students and their instructors,
who have pointed out aspects of the material where more or better explanations
and illustrations would be helpful. New topics in this edition include: default
modules for model design, limit cycles and chaos, parameter estimation in
Excel, model representations of gene regulation through transcription factors,
derivation of the Michaelis-Menten rate law from the original conceptual
model, different types of inhibition, hysteresis, a model of differentiation,
system adaptation to persistent signals, nonlinear nullclines, PBPK models, and
elementary modes.
I would like to thank three undergraduates from my classes who helped me with
the development of some of the new examples, namely Carla Kumbale, Kavya
Muddukumar, and Gautam Rangavajla. Quite a few other students have helped
me with the creation of new practice exercises, many of which are available on
the book’s support website. I also want to express my gratitude to David
Borrowdale, Katie Laurentiev, Georgina Lucas, Denise Schanck, and Summers
Scholl at Garland Science for shepherding this second edition through the
review and production process.
It is my hope that this new edition retains the appeal of the original and has
become even better through the alterations and twists it has taken, large and
small.
Eberhard Voit
Georgia Tech
2017
vi Instructor resources WebsIte
The author and publisher of A First Course in Systems Biology, Second Edition
gratefully acknowledge the contributions of the following reviewers in the
development of this book:
12.12 Modeling Heart Function and Failure 14.3 Design Principles 404
Based on Molecular Events 356 14.4 Operating Principles 406
Outlook for Physiological Multiscale Goal-Oriented Manipulations and Synthetic
Modeling 361 Design of Biological Systems 407
Exercises 362 14.5 Metabolic Engineering 407
References 365 14.6 Synthetic Biology 408
Further Reading 366 Case Studies of Synthetic Biological
Systems Designs 411
chapter 13: systems biology in Medicine 14.7 Elementary Mode Analysis in
and Drug Development 369 Metabolic Engineering 411
Are you Unique? 369 14.8 Drug Development 414
13.1 Biological Variability and Disease 369 14.9 Gene Circuits 415
13.2 Modeling Variability and Disease 370 The Future Has Begun 419
Personalized Medicine and Predictive Health 372 Exercises 419
13.3 Data Needs and Biomarkers 373 References 421
13.4 Personalizing Mathematical Models 374 Further Reading 423
The Drug Development Process 378
The Role of Systems Biology in chapter 15: emerging topics in
Drug Development 380 systems biology 425
13.5 Computational Target and Lead Emerging Applications 426
Identification 381 15.1 From Neurons to Brains 426
13.6 Receptor Dynamics 382 15.2 Complex Diseases, Inflammation,
13.7 Pharmacokinetic Modeling 385 and Trauma 428
13.8 Pathway Screening with Dynamic 15.3 Organisms and their Interactions
Models 390 with the Environment 432
13.9 Emerging Roles of Systems Biology Modeling Needs 435
in Drug Development 393 15.4 Multiscale Modeling 436
Exercises 394 15.5 A Data-Modeling Pipeline 437
References 395 Toward a Theory of Biology . . . or Several
Further Reading 396 Theories? 439
References 441
chapter 14: Design of biological systems 399 Further Reading 443
Natural Design of Biological Systems 400
14.1 The Search for Structural Patterns 400 Glossary 445
14.2 Network Motifs 402 Index 459
Biological Systems
1
When you have read this chapter, you should be able to:
• Describe the generic features of biological systems
• Explain the goals of systems biology
• Identify the complementary roles of reductionism and systems biology
• List those challenges of systems biology that cannot be solved with intuition
alone
• Assemble a “to-do” list for the field of systems biology
When we think of biological systems, our minds may immediately wander to the
Amazon rainforest, brimming with thousands of plants and animals that live with
each other, compete with each other, and depend on each other. We might think of
the incredible expanse of the world’s oceans, of colorful fish swimming through
coral reefs, nibbling on algae. Two-meter-high African termite mounds may come
to mind, with their huge colonies of individuals that have their specific roles and
whose lives are controlled by an intricate social structure (Figure 1.1). We may think
of an algae-covered pond with tadpoles and minnows that are about to restart yet
another life cycle.
These examples are indeed beautiful manifestations of some of the fascinating
systems nature has evolved. However, we don’t have to look that far to find biologi-
cal systems. Much, much smaller systems are in our own bodies and even within our
cells. Kidneys are waste-disposal systems. Mitochondria are energy-production sys-
tems. Ribosomes are intracellular machines that make proteins from amino acids.
Bacteria are amazingly complicated biological systems. Viruses interact with cells in
a well-controlled, systemic way. Even seemingly modest tasks often involve an
amazingly large number of processes that form complicated control systems
(Figure 1.2). The more we learn about the most basic processes of life, such as cell
division or the production of a metabolite, the more we have to marvel the incredi-
ble complexity of the systems that facilitate these processes. In our daily lives, we
usually take these systems for granted and assume that they function adequately,
and it is only when, for example, disease strikes or algal blooms kill fish that we
realize how complex biology really is and how damaging the failure of just a single
component can be.
We and our ancestors have been aware of biological systems since the beginning
of human existence. Human birth, development, health, disease, and death have
long been recognized as interwoven with those of plants and animals, and with the
environment. For our forebears, securing food required an understanding of sea-
sonal changes in the ecological systems of their surroundings. Even the earliest for-
ays into agriculture depended on detailed concepts and ideas of when and what to
2 Chapter 1: Biological Systems
plant, how and where to plant it, how many seeds to eat or to save for sowing, and
when to expect returns on the investment. Several thousand years ago, the Egyp-
tians managed to ferment sugars to alcohol and used the mash to bake bread. Early
pharmaceutical treatments of diseases certainly contained a good dose of supersti-
tion, and we are no longer convinced that rubbing on the spit of a toad during full
moon will cure warts, but the beginnings of pharmaceutical science in antiquity and
the Middle Ages also demonstrate a growing recognition that particular plant prod-
ucts can have significant and specific effects on the well-being or malfunctioning of
the systems within the human body.
In spite of our long history of dealing with biological systems, our mastery of
engineered systems far outstrips our capability to manipulate biological systems.
We send spaceships successfully to faraway places and predict correctly when they
will arrive and where they will land. We build skyscrapers exceeding by hundreds of
ABA PEPC
RCN1
NO
PLC PIP2 NAD+ ADPRc GTP GC InsPK Figure 1.2 Diagram of a complicated
PLD PC NADPH Atrboh
system of molecules that coordinate the
DAG InsP3 cADPR cGMP InsP6 RAC1 PA ROS response of plants to drought. While the
details are not important here, we can see
that a key hormone, called abscisic acid
CIS ABH1
ROP2 (ABA), triggers a cascade of reactions that
Actin ABI1 pHc ultimately promote the closure of stomata
and thereby reduce water evaporation [1].
ROP10 ERA1 CalM H+ ATPase
Even a narrowly defined response like this
Ca2+ ATPase closure process involves a complicated
Ca2+c KEV Depolar
control system that contains a multitude of
molecules and their interactions. In turn, this
AnionEM system is just one component within a much
KAP KOUT
larger, physiological stress response system
(cf. Figure 1.7). (From Saadatpour A, Albert I
& Albert A. J. Theor. Biol. 266 [2010] 641–656.
AtPP2C closure
With permission from Elsevier.)
BIOLOGICAL SYSTEMS 3
times the sizes of the biggest animals and plants. Our airplanes are faster, bigger,
and more robust against turbulence than the most skillful birds. Yet, we cannot cre-
ate new human cells or tissues from basic building blocks and we are seldom able to
cure diseases except with rather primitive methods like cutting into the body or kill-
ing a lot of healthy tissue in the process, hoping that the body will heal itself after-
wards. We can anticipate that our grandchildren will only shake their heads at such
medieval-sounding, draconian measures. We have learned to create improved
microorganisms, for instance for the bulk production of industrial alcohol or the
generation of pure amino acids, but the methods for doing so rely on bacterial
machinery that we do not fully understand and on artificially induced random
mutations rather than targeted design strategies.
Before we discuss the roots of the many challenges associated with understand-
ing and manipulating biological systems in a targeted fashion, and our problems
predicting what biological systems will do under yet-untested conditions, we should
ask whether the goal of a deeper understanding of biological systems is even worth
the effort. The answer is a resounding “Yes!” In fact, it is impossible even to imagine
the potential and scope of advances that might develop from biological systems
analyses. Just as nobody during the eighteenth century could foresee the ramifica-
tions of the Industrial Revolution or of electricity, the Biological Revolution will
usher in an entirely new world with incredible possibilities. Applications that are
already emerging on the horizon are personalized medical treatments with minimal
side effects, pills that will let the body regain control over a tumor that has run amok,
prevention and treatment of neurodegenerative diseases, and the creation of spare
organs from reprogrammed stem cells. A better understanding of ecological systems
will yield pest- and drought-resistant food sources, as well as means for restoring
polluted soil and water. It will help us understand why certain species are threat-
ened and what could be done effectively to counteract their decline. Deeper insights
into aquatic systems will lead to cleaner water and sustainable fisheries. Repro-
grammed microbes or nonliving systems composed of biological components will
dominate the production of chemical compounds from prescription drugs to large-
scale industrial organics, and might create energy sources without equal. Modified
viruses will become standard means for supplying cells with healthy proteins or
replacement genes. The rewards of discovering and characterizing the general prin-
ciples and the specifics of biological systems will truly be unlimited.
If it is possible to engineer very sophisticated machines and to predict exactly
what they will do, why are biological systems so different and difficult? One crucial
difference is that we have full control over engineered systems, but not over biologi-
cal systems. As a society, we collectively know all details of all parts of engineered
machines, because we made them. We know their properties and functions, and we
can explain how and why some engineer put a machine together in a particular
fashion. Furthermore, most engineered systems are modular, with each module
being designed for a unique, specific task. While these modules interact with each
other, they seldom have multiple roles in different parts of the system, in contrast to
biology and medicine, where, for instance, the same lipids can be components of
membranes and have complicated signaling functions, and where diseases are
often not restricted to a single organ or tissue, but may affect the immune system
and lead to changes in blood pressure and blood chemistry that secondarily cause
kidney and heart problems. A chemical refinery looks overwhelmingly complicated
to a layperson, but for an industrial engineer, every piece has a specific, well-defined
role within the refinery, and every piece or module has properties that were opti-
mized for this role. Moreover, should something go wrong, the machines and facto-
ries will have been equipped with sensors and warning signals pinpointing problems
as soon as they arise and allowing corrective action.
In contrast to dealing with sophisticated, well-characterized engineered sys-
tems, the analysis of biological systems requires investigations in the opposite direc-
tion. This type of investigation resembles the task of looking at an unknown machine
and predicting what it does (Figure 1.3). Adding to this challenge, all scientists col-
lectively know only a fraction of the components of biological systems, and the spe-
cific roles and interactions between these components are often obscure and
change over time. Even more than engineered systems, biological systems are full of
sensors and signals that indicate smooth running or ensuing problems, but in most
4 Chapter 1: Biological Systems
cases our experiments cannot directly perceive and measure these signals and we
can only indirectly deduce their existence and function. We observe organisms,
cells, or intracellular structures as if from a large distance and must deduce from
rather coarse observations how they might function or why they fail.
What exactly is it that makes biological systems so difficult to grasp? It is cer-
tainly not just size. Figure 1.4 shows two networks. One shows the vast highway
system of the continental United States, which covers several million miles of major
(A)
(B)
× 100
highways. It is a very large system, but it is not difficult to understand its function or
malfunction: if a highway is blocked, it does not take much ingenuity to figure out
how to circumvent the obstacle. The other network is a comparably tiny system: the
web of a diadem spider. While we can observe the process and pattern with which
Ms. Spider spins her web, we do not know which neurons in her brain are respon-
sible for different phases of the complicated web production process and how she
is able to produce the right chemicals for the spider silk, which in itself is a marvel
of material science, let alone how she manages to survive, multiply, and maybe
even devour her husband.
Biological systems often consist of large numbers of components, but they pose
an additional, formidable challenge to any analysis, because the processes that
govern them are not linear. This is a problem, because we are trained to think in
linear ways: if an investment of $100 leads to a return of $120, then an investment of
$10,000 leads to a return of $12,000. Biology is different. If we fertilize our roses with
1 tablespoon of fertilizer and the rose bushes produce 50 blossoms, a little bit more
fertilizer may increase the number of blossoms, but 100 tablespoons of fertilizer will
not produce 5000 blossoms but almost certainly kill the plants (Figure 1.5). Just a
small amount of additional sun exposure turns a tan into sunburn. Now imagine
that thousands of components, many of which we do not know, respond in such a
fashion, where a small input does not evoke any response, more input evokes a
physiological response, and a little bit more input causes the component to fail or
exhibit a totally different “stress” response. We will return to this issue later in this
and other chapters with specific examples.
species, along with descriptions of their leaves, berries, and roots, or their body
shapes, legs, and color patterns. These external descriptions were valuable, but did
not provide specific clues on how plants and animals function, why they live, and
why they die. Thus, the next logical step was to look inside—even if this required
stealing bodies from the cemetery under a full moon! Cutting bodies open revealed
an entirely new research frontier. What were all those distinct body parts and what
did they do? What were organs, muscles, and tendons composed of? Not surpris-
ingly, this line of investigation eventually led to the grand-challenge quest of discov-
ering and measuring all parts of a body, the parts of the parts (. . . of the parts), as well
as their roles in the normal physiology or pathology of cells, organs, and organisms.
The implicit assumption of this reductionist approach was that knowing the building
blocks of life would lead us to a comprehensive understanding of how life works.
If we fast-forward to the twenty-first century, have we succeeded and assembled
a complete parts catalog? Do we know the building blocks of life? The answer is a
combination of yes’s and no’s. The catalog is most certainly not complete, even for
relatively simple organisms. Yet, we have discovered and characterized genes, pro-
teins, and metabolites as the major building blocks. Scientists were jubilant when
the sequencing of the human genome in the early years of this millennium was
declared complete: we had identified the ultimate building blocks, our entire blue-
print. It turned out to consist of roughly three billion nucleotide pairs of DNA.
The sequencing of the human genome was without any doubt an incredible
achievement. Alas, there is much more to a human body than genes. So, the race for
building blocks extended to proteins and metabolites, toward individual gene varia-
tions and an assortment of molecules and processes affecting gene expression,
which changes in response to external and internal stimuli, during the day, and
throughout our lifetimes. As a direct consequence of these ongoing efforts, our parts
list continues to grow at a rapid pace: A parts catalog that started with a few organs
now contains over 20,000 human genes, many more genes from other organisms,
and hundreds of thousands of proteins and metabolites along with their variants. In
addition to merely looking at parts in isolation, we have begun to realize that most
biological components are affected and regulated by a variety of other components.
The expression of a gene may depend on several transcription factors, metabolites,
and a variety of small RNAs, as well as molecular, epigenetic attachments to its DNA
sequence. It is reasonable to expect that the list of processes within the body is much
larger than the number of components on our parts list. Biologists will not have to
worry about job security any time soon!
The large number of components and processes alone, however, is not the
only obstacle to understanding how cells and organisms function. After all, modern
computers can execute gazillions of operations within a second. Our billions of
telephones worldwide are functionally connected. We can make very accurate
REduCTIOnISM And SYSTEMS BIOLOGY 7
stomata SA JA ET
ETR1
NPR1 JAZ
EIN2
light ABA
defence
Glu HXK1
AUX GA
ABA
CK
ethylene
pollinator
H2O CO
2
O2
light
VOC
ORGAN AND PLANT GROWTH
temperature
pathogens
AUX CK
receptors enzymes S
CYCA DEL E2F RBR
other signaling proteins activation DP
CDKB CELL CYCLE
environmental interactions suppression
Figure 1.7 Stress responses are coordinated by systems at different levels of organization (cf. Figure 1.2). At the physiological level, the stress
response system in plants includes changes at the cellular, organ, and whole-plant levels and also affects interactions of the plant with other species.
(From Keurentjes JJB, Angenent GC, Dicke M, et al. Trends Plant Sci. 16 [2011] 183–190. With permission from Elsevier.)
8 Chapter 1: Biological Systems
quite fair, because, in addition to their large number, the components of a cell are not
all the same, which drastically complicates matters. Furthermore, as mentioned ear-
lier, the processes with which the components interact are nonlinear, and this per-
mits an enormous repertoire of distinctly different behaviors with which an organism
can respond to a perturbation.
Here, X, Y, and Z are concentrations, E is the enzyme activity, and a, b, and c are rate 1.0
constants that respectively represent how fast X is converted into Y, how fast Y is concentration X Z
converted into Z, and how quickly material from the metabolite pool Z leaves the Y
0.5
system. The dotted quantities on the left of the equal signs are differentials that
describe the change in each variable over time, but we need not worry about them
at this point. In fact, we hardly have to analyze these equations mathematically to 0
0 15 30
get an idea of what will happen if we change the input, because intuition tells us that time
any increase in Input should lead to a corresponding rise in the concentrations of
the intermediates X, Y, and Z, whereas a decrease in Input should result in smaller Figure 1.9 Simulations with the system
values of X, Y, and Z. The increases or decreases in X, Y, and Z will not necessarily be in (1.1) confirm our intuition: X, Y, and
exactly of the same extent as the change in Input, but the direction of the change Z reflect changes in Input. For instance,
should be the same. The mathematical solution of the system in (1.1) confirms this reducing Input in (1.1) to 75% at time
intuition. For instance, if we reduce Input from 1 to 0.75, the levels of X, Y, and Z 10 (arrow) leads to permanent decreases
decrease, one after another, from their initial value of 1 to 0.5625 (Figure 1.9). in X, Y, and Z.
Now suppose that Z is a signaling molecule, such as a hormone or a phospho-
lipid, that activates a transcription factor TF that facilitates the up-regulation of a
gene G that codes for the enzyme catalyzing the conversion of X into Y (Figure 1.10).
The simple linear pathway is now part of a functional loop. The organization of this
loop is easy to grasp, but what is its effect? Intuition might lead us to believe that the
positive-feedback loop should increase the level of enzyme E, which would result in E G TF
more Y, more Z, and even more E, which would result in even more Y and Z. Would
the concentrations in the system grow without end? Can we be sure about this pre-
diction? Would an unending expansion be reasonable? What will happen if we Input X Y Z
increase or decrease the input as before?
The overall answer will be surprising: the information given so far does not allow
us to predict particular responses with any degree of reliability. Instead, the answer Figure 1.10 Even simple systems may
depends on the numerical specifications of the system. This is bad news for the not allow us to make reliable predictions
unaided human mind, because we are simply not able to assess the numerical con- regarding their responses to stimuli.
Here, the linear pathway from Figure 1.8 is
sequences of slight changes in a system, even if we can easily grasp the logic of a
embedded into a functional loop consisting
system as in Figure 1.10. of a transcription factor TF and a gene G that
To get a feel for the system, one can compute a few examples with an expanded codes for enzyme E. As described in the text,
model that accounts for the new variables (for details, see [3]). Here, the results are the responses to changes in Input are no
more important than the technical details. If the effect of Z on TF is weak, the longer obvious.
EVEn SIMPLE SYSTEMS CAn COnFuSE uS 9
response to a decrease in Input is essentially the same as in Figure 1.9. This is not too
surprising, because the systems in this case are very similar. However, if the effect of
Z on TF is stronger, the concentrations in the system start to oscillate, and after a
while these oscillations dampen away (Figure 1.11A). This behavior was not easy to
predict. Interestingly, if the effect is further increased, the system enters a stable
oscillation pattern that does not cease unless the system input is changed again
(Figure 1.11B).
The hand-waving explanation of these results is that the increased enzyme activ-
ity leads to a depletion of X. A reduced level of X leads to lower levels of Y and Z,
which in turn lead to a reduced effect on TF, G, and ultimately E. Depending on the
numerical characteristics, the ups and downs in X may not be noticeable, they may
be damped and disappear, or they may persist until another change is introduced.
Intriguingly, even if we know that these alternative responses are possible, the
unaided human mind is not equipped to integrate the numerical features of the
model in such a way that we can predict which system response will ensue for a
specific setting of parameters. A computational model, in contrast, reveals the
answer in a fraction of a second.
The specific details of the example are not as important as the take-home mes-
sage: If a system contains regulatory signals that form functional loops, we can no
longer rely on our intuition for making reliable predictions. Alas, essentially all real-
istic systems in biology are regulated—and not just with one, but with many control
loops. This leads to the direct and sobering deduction that intuition is not sufficient
and that we instead need to utilize computational models to figure out how even
small systems work and why they might show distinctly different responses or even
fail, depending on the conditions under which they operate.
The previous sections have taught us that biological systems contain large num-
bers of different types of components that interact in potentially complicated ways
and are controlled by regulatory signals. What else is special about biological sys-
tems? Many answers could be given, some of which are discussed throughout this
book. For instance, two biological components are seldom 100% the same. They vary
from one organism to the next and change over time. Sometimes these variations are
inconsequential, at other times they lead to early aging and disease. In fact, most
(A)
2
X
concentration
TF, E, G
Z
Y
0
0 50 100
time
(B)
5.0
diseases do not have a single cause, but are the consequence of an unfortunate com-
bination of slight alterations in many components. Another feature that complicates
intuition is the delay in many responses to stimuli. Such delays may be of the order of
seconds, hours, or years, but they require the analyst to study not merely the present
state of a biological system but also its history. For instance, recovery from a severe
infection depends greatly on the preconditioning of the organism, which is the col-
lective result of earlier infections and the body’s responses [4].
Finally, it should be mentioned that different parts of biological systems may
simultaneously operate at different scales, with respect to both time and space.
These scales make some aspects of their analysis easier and some harder. Let’s begin
with the temporal scale. We know that biology at the most basic level is governed by
physical and chemical processes. These occur on timescales of the order of millisec-
onds, if not faster. Biochemical processes usually run on a scale of seconds to min-
utes. Under favorable conditions, bacteria divide every 20–30 minutes. Our human
lifespan extends to maybe 120 years, evolution can happen at the genetic level with
lightning speed, for instance, when radiation causes a mutation, while the emer-
gence of an entirely new species may take thousands or even millions of years. On
one hand, the drastically different timescales make analyses complicated, because
we simply cannot account for rapid changes in all molecules of an organism over an
extended period of time. As an example, it is impossible to study aging by monitor-
ing an organism’s molecular state every second or minute. On the other hand, the
differences in timescales justify a very valuable modeling “trick” [5, Chapter 5]. If we
are interested in understanding some biochemical process, such as the generation
of energy in the form of adenosine triphosphate (ATP) by means of the conversion
of glucose into pyruvate, we can assume that developmental and evolutionary
changes are so slow in comparison that they do not change during ATP production.
Similarly, if we study the phylogenetic family tree of species, the biochemical pro-
cesses in an individual organism are comparatively so fast that their details become
irrelevant. Thus, by focusing on just the most relevant timescale and ignoring much
faster and much slower processes, any modeling effort is dramatically simplified.
Biology also happens on many spatial scales. All processes have a molecular
component, and their size scale is therefore of the order of ångströms and nanome-
ters. If we consider a cell as the basic unit of life, we are dealing with a spatial scale
of micrometers to millimeters, with some exceptions such as cotton “fiber” cells
reaching the length of a few centimeters [6] and the afferent axons of nerve cells in
giraffes, reaching from toe to neck, extending to 5 meters [7, p. 14]. The sizes of typi-
cal cells are dwarfed by higher plants and animals and by ecosystems such as our
oceans, which may cover thousands of square kilometers. As with the different tem-
poral scales, and using analogous arguments, models of biological systems often
focus on one or two spatial scales at a time [5]. Nonetheless, such simplifications are
not always applicable, and some processes, such as aging and algal blooms, may
require the simultaneous consideration of several temporal and spatial scales. Such
multiscale assessments are often very complicated and constitute a challenging
frontier of current research (see Chapter 15).
WHY nOW?
Many of the features of biological systems have been known for quite a while, and,
similarly, many concepts and methods of systems biology have their roots in its
well-established parent disciplines, including physiology, molecular biology, bio-
chemistry, mathematics, engineering, and computer science [8–11]. In fact, it has
been suggested that the nineteenth-century scientist Claude Bernard might be con-
sidered the first systems biologist, since he proclaimed that the “application of
mathematics to natural phenomena is the aim of all science, because the expression
of the laws of phenomena should always be mathematical” [12, 13]. A century later,
Ludwig von Bertalanffy reviewed in a book his three decades of attempting to con-
vince biologists of the systemic nature of living organisms [14, 15]. At the same time,
Mihajlo Mesarović used the term “Systems Biology” and declared that “real
advance . . . will come about only when biologists start asking questions which are
based on systems-theoretic concepts” [16]. The same year, a book review in Science
WHY nOW? 11
envisioned “. . . a field of systems biology with its own identity and in its own right”
[17]. A few years later, Michael Savageau proposed an agenda for studying biologi-
cal systems with mathematical and computational means [5].
In spite of these efforts, systems biology did not enter the mainstream for several
more decades. Biology kept its distance from mathematics, computer science, and
engineering, primarily because biological phenomena were seen as too complicated
for rigorous mathematical analysis and mathematics was considered applicable only
to very small systems of little biological relevance. The engineering of biological sys-
tems from scratch was impossible, and the budding field of computer science con-
tributed to biology not much more than rudimentary data management.
So, why has systems biology all of the sudden moved to the fore? Any good detec-
tive will know the answer: motive and opportunity. The motive lies in the realization
that reductionist thinking and experimentation alone are not sufficient if complex
systems are involved. Reductionist experiments are very good in generating detailed
information regarding specific components or processes of a system, but they often
lack the ability to characterize, explain, or predict emergent properties that cannot
be found in the parts of the system but only in their web of interactions. For instance,
the emergence of oscillations in the example system represented by the equations
in (1.1) cannot be credited to a single component of the system but is a function of
its overall organization. Although we had complete knowledge of all details of the
model pathway, it was very difficult to foresee its capacity either to saturate or oscil-
late in a damped or stable fashion. Biology is full of such examples.
A few years ago, Hirotada Mori’s laboratory completed the assembly of a com-
plete catalogue of single mutants in the bacterium Escherichia coli [18]. Yet, the
scientific community is still not able to foresee which genes the bacterium will up-
or down-regulate in response to new environmental conditions. Another very chal-
lenging example of emergent system properties is the central nervous system. Even
though we understand quite well how action potentials are generated and propa-
gated in individual neurons, we do not know how information flows, how memory
works, and how diseases affect the normal functioning of the brain. It is not even
clear how information in the brain is represented (see also Chapter 15). Thus, while
reductionist biology has been extremely successful and will without any doubt
continue to be the major driving force for future discovery, many biologists have
come to recognize that the detailed pieces of information resulting from this
approach need to be complemented with new methods of system integration and
reconstruction [19].
The opportunity for systems biology is the result of the recent confluence and
synergism of three scientific frontiers. The first is of course the rapid and vast accu-
mulation of detailed biological information at the physiological, cellular, molecular,
and submolecular levels. These targeted investigations of specific phenomena are
accompanied by large-scale, high-throughput studies that were entirely infeasible
just a couple of decades ago. They include quantification of genome-wide expres-
sion patterns, simultaneous identification of large arrays of expressed proteins,
comprehensive profiling of cellular metabolites, characterization of networks of
molecular interactions, global assessments of immune systems, and functional
scans of nervous systems and the human brain. These exciting techniques are gen-
erating unprecedented amounts of high-quality data that are awaiting systemic
interpretation and integration (Figure 1.12).
The second frontier is the result of ingenuity and innovation in engineering,
chemistry, and material sciences, which have begun to provide us with a growing
array of technologies for probing, sensing, imaging, and measuring biological sys-
tems that are at once very detailed, extremely specific, and usable in vivo. Many
tools supporting these methods are in the process of being miniaturized, in some
cases down to the nanoscale of molecules, which allows diagnoses with minute
amounts of biological materials and one day maybe biopsies of individual, living
cells. Devices at this scale will allow the insertion of sensing and disease treatment
devices into the human body in an essentially noninvasive and harmless fashion
[20–22]. Bioengineering and robotics are beginning to render it possible to measure
hundreds or thousands of biomarkers from a single drop of blood. It is even becom-
ing feasible to use molecular structures, prefabricated by nature, for new purposes
in medicine, drug delivery, and biotechnology (Figure 1.13).