Synthetic biology is the study of artificial living organisms, re-engineered organisms, and engineered parts for existing organisms. The idea is to bring engineering tools and methodologies to the subject of biology, much as electrical engineering brings these tools to physics, with the goal of designing novel biological systems. These novel systems, new life forms really, may someday afford humanity a new level control over the processes in our bodies, in our crops, in landfills, compost heaps, and the ocean waters: Instead of the current brute force attempts to control the living world with machines and harsh chemicals, synthetic biology promises control at the level of individual molecules using engineered living cells as tiny chemical processing plants.
Indeed, biology is more than just chemical processing. It is information processing. The instructions for how to grow and regulate a human being are encoded in the DNA inside each of the cells in your body. Growth and development follows precise, distributed, robust algorithms. The immune system can adapt itself to pathogens it has never seen. A predator can track its prey and outsmart it. Our brains can do research and write books. Life is computation. The molecules and structures inside cells are the tools that nature uses to implement life's computation, just as the silicon and wires inside your computer are our way to implement computations. Thus to some extent, biology (synthetic or otherwise) ought to be seen more abstractly than our high-school biology books would have us see it. Our job as living system engineers is to define the abstractions, the composable modules, the signal carriers, the simplificatons, the programming languges, and the debugging tools for living systems.
Below is an overview of our projects in synthetic biology. The field is young, and so the projects are focused on very detailed mechanisms. It will be a while before we can make new living systems from scratch, but working on fundamental design principles will enable it to happen. Please feel free to email Prof. Klavins if you would like more details.
Control of Stochasticity in Gene Expression
When we compare the behaviors of individual isogenic cells in a microscope, we see substantial variation in gene expression. The variability, or noise, comes from a variety of factors. We are mainly interested in noise due to copy-number variation. That is, there may be between 0 and 100 copies of a particular protein in a cell. Thus, if one more gets expressed, or one degrades, between a 0% and 100% change in the overall expression will occur. This fact creates many challenges for engineering biology, and new methods of understanding and coping with such noise must be developed before large scale circuits can be constructed in cells. The Klavins lab is interested in how and to what degree noise in single cells is controlled by the cell, and how robust and predictable behaviors can be synthesized for cells, given that noise cannot be avoided. To that end, we focus on noise in simple genetic regulatory networks and attempt to understand how noise properties depend on the architecture of the circuits. Of particular interests are situations where noise is actually useful. For example, many algorithms in computer science rely on randomness to work, and certain tasks can even be proved to be impossible without random number generators. Perhaps living systems use random number generators the same way.
Tuning Gene Expression
An important feature of electrical circuits is the placement of tuning knobs: variable elements that can be tuned to account for the uncertainty in the rest of the components of the circuit. By tuning a circuit appropriate, the correct behavior, the best performance, and the most robustness can be obtained. On the other hand, an incorrectly tuned circuit might not do anything at all. We are developing several methods for tuning genetic regulatory networks based on modifying transcription and/or translation rates, binding affinities, etc. The long term goals are (a) to understand how and where to build tunability into biomolecular systems and (b) to devise experimental methods in which a circuit architecture can be tuned into place quickly and easily. For now, our experimental systems are simple circuits in E. coli and auxin processing circuits in yeast (see below).
Auxin Signal Processing
The auxin pathway is central to nearly every aspect of higher plant life and evolution. We are reverse engineering the auxin pathway using Saccharomyces cerevisiae as a testbed. The fact that different tissues and different plants use different combinations of Auxin Response Factors (ARFs) and their inhibitors (Aux/IAAs) for a huge array of purposes suggests that the auxin pathway is indeed reprogrammable.
Together with the Nemhauser Lab, we are rebuilding the auxin pathway so that we can learn to program with it. Our goals are to improve our understanding of individual parts of the pathway and to build entirely new systems capable of producing novel behaviors. In the short term, we intend to reproduce the behaviors of simple multi-celled organisms or tissues in a population of engineered yeast cells. Next, we expect to port a complete set of signal processing components that could be used in essentially any organism as a modular auxin sensor.
Ultimately perhaps the most interesting aspect of auxin processing is how auxin is used in development in plants. Development requires, in addition to the pathway we are investigating, auxin transport proteins, and auxin synthesis, degradation, and sequestration. These systems together allow the development plant to coordinate the growth patterns of all the cells in the plant to create the ultimate shape of the plant. Porting all of these systems to yeast could allow us to reprogram non-plant cells to to develop into patterns as well. Understanding development by recapitulating it in a single celled organism is a truly motivating engineering goal (to us, anyway).
Interacting Nucleic Acid Circuits
DNA is a remarkable molecule, not in the least because it is so easy to program. Take a couple of strands that are complimentary and put them in a test tube, and you get the reaction . Through clever manipulation, you can in fact implement any chemical reaction you want (see Prof. Georg Seelig's work, for example). You might also try to implement other formalisms besides abstracts chemical reactions. The Klavins group is in particular interested in implementing feedback control methods. This requires that we understand how to represent, with nucleic acids, arbitrary signals, and how to operate on those signals to produce new signals. We also have to understand how to wire up smaller systems into bigger systems -- that is, to compose systems.
We explore DNA devices with two different experimental frameworks. The first is DNA only. It turns out to be possible to design a set of nucleic acids and nucleic acid complexes that implement any linear I/O system, assuming certain properties on the concentrations of "fuel" strands and complexes. See Kevin Oishi's paper on the subject. The second system consists of "genelets" made from DNA, RNA signals, RNA polymerase, and Ribonuclease H. The framework was invented by Winfree's group at Caltech. In principle, one can build arbitrary transcriptional regulatory networks by making the RNA output of one genelet either repress or activate the production of RNA from another genelet. We are particularly interested in feedback circuits, such as the one shown to the right. The idea is to produce an RNA output that is insensitive to loads from downstream devices. If the load increases, the circuit produces more RNA to compensate. For both of these systems, we are developing design theories and testing designs in experiments.
Nucleic acid circuits are currently rather idealized systems with no direct applications. They provide a framework in which our fundamental understanding of can be tested. However, nucleic acid circuits may be very useful in a variety of settings. For example, low cost biosensors might amplify an RNA sample that might indicate a diseased state. Even more exciting, learning to program nucleic acid circuits in test tubes, might teach us how to program similar circuits in cells that use siRNAs to modulate gene expression -- which could lead to novel gene therapies and other applications.
State Estimation in Gene Expression
Decentralized Robot Systems
Projects in this area will end in 2011.
Verification and Validation of Complex Systems
Projects in this area will end in 2011.
- NSF (co-PI with Jeff Tabor, Ben Kerr, Oleg Igoshin, and Georg Seelig): Harnessing Intercellular Signaling to Engineer Pattern Formation, 2011-2014.
- PGAFF (co-PI with Jennifer Nemhauser): Reprogramming Cells with Plant-Derived Signalling Pathways, 2011-2014.
- NSF (PI with David Thorsley (co-PI)): Estimation & Observation of Stochastic Biochemical Networks, 2010-2013.
- NSF (co-PI): The Molecular Programming Project, 2008-2013.
- NSF (co-PI), EFRI: Controlling the Autonomously Reconfiguring Factory, 2008-2011.
- AFOSR (co-PI), MURI: Verification and Validation of Distributed Networked Systems, 2006-2011.
- Microsoft (PI) A. Richard Newton Breakthrough Research Award: The Stochastic Model Builder Applied to Single Cell Kinetics, 2008-2009.
- UW College of Engineering (PI): Interdisciplinary Undergraduate Sequence in Systems and Synthetic Biology, 2008-2009 (with H. Sauro).
- UW College of Engineering (PI): Curriculum Development for Synthetic Biology, 2007-2008.
- NSF (PI): CAREER: Programmed Robotic Self-Assembly, 2004-2009 (apparently, my career is over!)
- NSF (co-PI): 3D Self-Assembly, 2004-2008.
- DARPA (co-PI): 3D Directed Self-Assembly, 2004-2005.
- UW Royalty Research Fund (PI): A Synthesis Method for DNA Machines, 2004-2005.
- NSF (co-PI): A Computing Lab for Integrated Teaching of Systems Courses in Electrical Engineering, 2005-2007.
- In 2007, ICOS kindly donated two liquid handlers and a CRS robotic arm fitted for laboratory automation. Thanks!