Abstract:
A central problem in "post-genomic biology" is reverse and forward
engineering the dynamics and control of intracellular networks of genes and
proteins. These interact through intercellular chemical and mechanical
signaling to direct both development and regulate an organism's response to
its environment. Similar challenges exist in understanding the more global
regulatory strategies that maintain organism and ecosystem homeostasis.
These complex networks can overwhelm intuition and informal models and thus
modeling and simulation methods are playing an increasingly central role,
inspired by computer-aided engineering and scientific computation. One
important lesson can and must be drawn from the history of these areas:
brute-force modeling and computation has no chance of succeeding for systems
with the type of complexity of a biological cell, let alone an organism or
ecosystem.


Thus the success of post-genomic biology will be contingent upon the
development of new algorithms and methodologies guided by rigorous
mathematical theory. A corresponding challenge in "better, cheaper,
faster" engineering is to create robust, reliable systems using more
virtual and less physical prototyping, and greater component reuse. Hard
problems include multiscale integration of electrical, mechanical, and
chemical subsystems. Furthermore, the verification of complex engineering
systems with embedded software closely parallels the robustness analysis of
complex biological developmental and regulatory pathways controlled by
"embedded" computational networks of genes and proteins.


Two great abstractions of 20th century engineering are that control,
communications, and computing could be developed 1) largely separately from
each other, and 2) independently of the details of physical substrates.
This held both in practical applications and academic research, and
facilitated a massive and parallel technology development. Popular
technological visions employing MEMS and ubiquitous, embedded, distributed
networking, controls and computing will stretch these abstractions past
their breaking point. While biological cellular processes are far more
distributed, stochastic, and heterogeneous than their engineering
counterparts, in both cases, these form merely the components of complex
control, communications, and computing systems. Biology not only integrates
these functions but also builds them directly at the molecular level. Since
the research needs for systems engineering and systems biology are
converging, it is fortunate that the mathematical theory and software
infrastructure to address these needs is finally an achievable goal.
Central to this theory is a growing understanding of the "design principles"
of complex networks, and the role of protocols, modularity, and
interconnection in creating robust, evolvable systems.