Biomedical Engineering Reference
In-Depth Information
helps uncover new relationships. Early attempts at building biological models
focused primarily on metabolic pathways because of the availability of a rich
inventory of qualitative and quantitative data. With the emergence of new in-
formation, researchers moved from a data-modeling to a process-modeling ap-
proach (36) using kinetic equations and rate laws. Some of the earliest papers in
biochemical pathway modeling were published during the 1960s and 1970s
(8,14,25-27). It was quite challenging to solve algebraic equations manually,
thus restricting the "model bandwidth" to only a few equations. During the
1980s and 1990s, the availability of massive computational power tilted the bal-
ance in favor of more powerful modeling strategies (9,10). However, despite the
impressive strides in biological computer modeling, a few constraints restricted
its full-scale growth. Some of the constraints were (a) a lack of high throughput
and high-quality data, (b) understanding of complexity, and (c) difficulty in ex-
perimentally validating computer models. In 1997 a major milestone in biologi-
cal modeling was reached when a virtual
Mycoplasma genitalium
with 127
genes was created (34), thereby signaling the arrival of an era of credible in-
silico modeling. Over the last few years, using sound theoretical and experimen-
tal data, biochemical and gene expression models have been published, pushing
the field of in-silico modeling to new levels. Excellent reviews have appeared on
this subject recently (3,12,16,17,23,35,39).
3.
MODELING AND SIMULATION
Biochemical systems come with a variety of features: forward reactions,
reverse reactions, feedback loops, redundancy, stability, and modularity. The
challenge is to quantitatively represent each of these features and integrate them
into the model. A number of parallels between biological systems and engineer-
ing systems have prompted researchers to adopt reverse-engineering approaches,
using well-established concepts from physics. Some of the traits that biological
and engineering systems share include: rapid communication and response, ac-
curate error detection and correction, fuzzy control, amplification, adaptation,
and robustness. However, biological systems are open (i.e., interact with the
environment, thereby providing an unlimited supply of building blocks for nu-
cleic acids and proteins), nonlinear (a well-defined input does not always lead to
the predicted output) (see this volume, Part II, chapter 2, by Socolar), and ex-
hibit an emergent property (system behavior cannot be explained by individual
components).
3.1. Why Do We Need Modeling?
1. Every aspect of biological phenomena cannot be easily captured by ex-
periments alone. To answer complex "what if" questions, one needs novel tools
and strategies to supplement the wet bench approach.
