Keynote speakers

Prof. Richard B. Fair
Department of Electrical and Computer Engineering, Duke University, Durham, NC, USA

Progress in Reconfigurable Microfluidic Systems for Evolvable BioChips

Microfluidic systems serve as a key link between evolutionary algorithms run on electronic hardware and chemical and biological applications performed on microscale biochip hardware. Historically, microfluidic devices have been highly application specific, lacking reconfigurability and programmability. Such "fixed data path" devices are the current norm today for commercial biochips. Nevertheless, the need for reconfigurable microfluidic devices was recognized over 15 years ago. What was missing in the implementation of these devices was the equivalent of switches (transistors in FPGAs) that would allow for dynamic switchable fluidic connections and processes.

However, two innovations have occurred since then: 1) pneumatic monolithic microvalves that can be used to fabricate large-scale microfluidic integrated systems, thus increasing the processing power of a network by allowing complex fluid manipulations with a minimal number of controlled inputs, and 2) digital microfluidic devices that enable versatile, reconfigurable chip architectures and that are capable of accommodating and adapting to multiple applications. Establishing a development path (technology and commercialization) for programmable microfluidics that parallels the development of FPGAs requires the definition of architectural and execution concepts for assembling microfluidic devices into networks that perform fluidic operations in support of a diverse set of applications. Indeed a hierarchical integrated microfluidic design approach is described to facilitate scalable design for many biomedical applications. It is demonstrated that a wide diversity in biomedical applications can be parsed into manageable components and assembled into an architecture proving the advantages of being programmable, reconfigurable, and reusable. This capability opens the possibility of handling all of the protocols that a given laboratory or a class of applications would require. And, it provides a path toward realizing evolvable biochip hardware. Progress in implementing this capability is demonstrated with examples of a complete set of programmable elemental fluidic components that support all of the required fluidic operations in a biochip.

Prof. Jordan Pollack
Computer Science Department, Volen Center for Complex Systems, Brandeis University, USA

A new approach to machine embryogenesis

I will review several generations of machines which were automatically designed using co-evolutionary techniques in my laboratory. While the goal was also fully automated construction by robot factories, in each case, final assembly has been supplemented with varying degrees of help from human minds and fingers.

I originally felt that these manual operations could be automated for a price, but now I believe that the assembly of complex mechanism (e.g. not pure logic or photo-lithographic) is a much harder problem than anyone suspected.

In Nature, the embryogenesis process proceeds from a single fertilized egg through division, migration, and specialization. The egg is an assembly factory, without any human supervisors, without any brain, which produces an exquisite and custom product with 10 billion moving parts in only 9 months!

Yet every assembly factory in the real world depends critically on human minds both as labor as well as supervision to monitor, correct and repair ongoing processes. Where is the mind inside the fertilized egg? Even proponents of Intelligent Design would be hard pressed to defend the existence of an omniscient "Intelligent Factory Foreman" who supervises every embryo developing in the world simultaneously, deciding which creatures live or die.

Although a lot is known about development, we have a long way to go from theories of pattern formation towards understanding the supervisory intelligence of a process which assembles complex biological form.

In order to solve our robot-embryogenesis problem, we have replaced the idea of a perfect factory with one subject to noise and error. After some work in evolving assembly plans which could overcome errors through redundancy, we noticed an isomorphism between noisy robotic assembly and certain games which fit under Bellman equation. This leads to a new theory I will sketch for bio-mimetic self-construction: DNA as a "blind chessmaster" which continuously optimizes assembly processes in a game against Nature.

Prof. Jaroslav Flegr
Faculty of Science, Charles University in Prague, Czech Republic

Postneodarwinistic theories of biological evolution - From selfish gene to frozen evolution

The mechanism of natural selection can lead to the evolution of adaptive traits only amongst asexually reproducing organisms. Amongst organisms that reproduce sexually, the genotype of the individual is formed in each generation through the random mixing of the genes of the two parents and thus the biological fitness of individuals is not inherited. The theory of the selfish gene was proposed in the 1970's to explain the functioning of adaptive evolution amongst sexually reproducing organisms. According to this theory, individuals do not compete for the greatest biological fitness in the framework of the population during evolution, but rather the alleles compete in the framework of a single locus for the ability to transfer the greatest number of their copies to the gene pool of the next generation. However, this theory neglects the fact that the effect of the individual alleles on the phenotype and the effect of the individual phenotype traits on the biological fitness of individuals depend on the other alleles that are present in the genotype of the individual.

The theory of evolutionarily stable strategies indicates that, under these conditions, selection pressures cannot lead to long-term changes in the phenotypes of organisms, but only to deflection of the frequency of the individual alleles from equilibrium. The greater this deflection, the more the gene pool resists this pressure and, after cessation of the selection pressure, the frequency of the alleles spontaneously returns to the original values.

The theory of frozen plasticity, published in 1998, shows that sexually reproducing species can respond evolutionarily to selection pressures (they are evolutionarily plastic) only when the members of the particular species are genetically uniform, i.e. after splitting off and subsequent rapid multiplication of part of the population of the original species. Following a short period of time, estimated on the basis of paleontological data to correspond to 1-2% of the duration of the species, genetic polymorphism accumulates in the gene pool and thus, in each generation, the new mutations are in the presence of different alleles - the species ceases to behave in an evolutionarily plastic manner and begins to be evolutionarily elastic. It then exists in this state until such time as such changes accumulate in the environment that the evolutionarily frozen species becomes extinct.