Thoughts on the Biology / EECS Relationship

Tom Knight

1/23/03
Additions 5/26/05/

In the 1930's, electrical engineering transitioned from a branch of physics into a separate discipline. Key to this notion was the concept of "components" which could be analyzed, designed, characterized, and manufactured independently -- together with the
remarkable ability to hierarchically interconnect those modular components into a bewildering array of devices. Moving from physics, which told us be basic laws of electromagnetism, to engineering, which abstracted those laws into simpler, understandable form, was the key contribution of the intellectual pioneers of the discipline such as Armstrong, Black, Blumlein, and Guillemin.

The transition from electrical to electronic systems happened with an understanding of signals and their representations -- we moved from thinking primarily about power systems to thinking primarily about the information represented by electrical signals. Shannon, digital representations, and the entire development of computing technology
are the direct result.

In most respects the association of electrical engineering with electromagnetism is now almost incidental. We have become complexity and information engineers, rather than experts in E&M. Few other disciplines design, build, and debug systems as complex as a modern computer system, either from the perspective of hardware, with billions of components, or with software, with millions of lines of code.

We are uniquely positioned to address the issues of understanding, modeling, building, and debugging complex systems.

Yet, I see that the focus of our relationships to biology are driven by a much narrower perspective. Often they are driven by the notion that we should build technology in service to the biological community. There is absolutely nothing wrong with this -- in fact,
there is much right -- but we can do so much more.

In addition to developing the instrumentation and analytical tools for biological systems, we have the opportunity to revolutionize biology, and to reinvigorate our own discipline.

To me, three major themes appear to be emerging:

First, the lessons learned by evolution in the development, maintenance, and organization of complex systems from large numbers of unreliable components are becoming central to managing the complex information systems of the current generation. We need to learn lessions from biological systems about how to respond to local failure while maintaining global function. This is is a necessary part of today's engineering in a networked environment (or with current generation VLSI technology).

Second, we have an opportunity to use our complexity and information management tools to understand, modularize, abstract, and understand biological systems. In the same way that we simplified and abstracted components from physics to allow us to build billion component processors, we can and will modularize, abstract, and understand biological components with the explicit goal of constructing artificial biochemical and biological systems. I believe no other discipline can effectively do this.

Third, we have an opportunity to reinvent our technological substrate. Molecular electronics is poised to push silicon into the background, but it needs plausible fabrication mechanisms for complex systems. The replacement of lithography, which can fabricate billions of components in parallel, is a particular challenge. Patterned molecular scale structures made with chemically precise biopolymers are, I believe, the only plausible choice. Biology can make these, if we can control the biology.

It is by now a cliche to say that biology will be the discipline of the century. But I would argue that the real revolution will come not from the medical and scientific aspects, but through exerting control over the signals, the modularity, the interfaces, and the complexity of these systems -- aspects which we, as a discipline, are uniquely
capable of addressing.

A new molecular scale manufacturing technology will emerge, driven by the unique self-reproducing capabilites of biological systems. We don't need to invent nanotechnology -- we already have it. We just need to tame it, understand it, and control it. Along the way, we will push the manufacturing capabilites of our own discipline, electronics, to currently unimaginable densities, speeds, and performance.

My own work impacts in this area in two ways:

First, I am developing the concept of biological parts, with defined engineering behavior. Setting in place the categories of parts, the units in which their properties are measured, and the standardized methods of measurement is the first order of business. Many of the
existing parts are slight modifications of natural components, carefully selected for their engineering utility. Others are beginning to be crafted de novo with specific engineering goals in mind. The collection of parts, which I term BioBricks, has been designed with a unique hierarchical assembly technique which I am standardizing and automating. With the recent help of Randy Rettberg and Drew Endy, this collection of parts has grown include about five hundred components, and is well documented in the web site of the MIT Registry of Standard Biological Parts, at http://parts.mit.edu. Thirteen universities in four countries are using these parts this summer as part of the intercollegiate genetically engineered machines challenge, as well as in more routine engineering efforts.

Second, I am engaged in a long term effort to construct a simple living cell as an engineering vehicle for living systems. Over the past six years, I have intensively studied the class of organisms known as Mycoplasmas, the smallest and simplest of free living cells. Most of these organisms are plant, animal, or human pathogens, and are thus unsuitable for routine engineering use. I have selected the organism Mesoplasma florum as a target organism because it is non-pathogenic, and does not grow at 37C. After preliminary analysis and sequencing of approximately 17% or the genome in my laboratory, shotgun sequencing and assembly was done in collaboration with Whitehead. Together with Greg Fournier working in my laboratory, I fully annotated the 793Kb genome, and am preparing a very detailed paper analyzing the sequence. I plan to redesign the genome of the species from the ground up, with the goal of creating a very well understood, carefully crafted organism suitable as a simple living substrate for the nascent engineering discipline of synthetic biology.

A Scientist discovers that which exists. An Engineer creates that which never was.
-- Theodore von Karman