I was educated as an engineer and physicist. I now work on theoretical models of biological processes, mostly but not exclusively at the molecular level.
My interests in biology have centered from the beginning on the theoretical implications of the complex relations between genotype and phenotype, and their consequences for morphogenesis and evolution (Kerszberg and Agur, 1987). From 1993 onwards, this basic motivation has drawn me increasingly towards the integrative aspects of biology.
I have invested a large effort in the construction of ctrl-Dev (control Dev), an experimental computer system adapted to the description and simulation of genetical, molecular and biomechanical processes in cells and groups of cells, i.e. in tissues. The originality of the system is, first, that it integrates within a single framework cell motility and genetic interactions; and second, that it takes into account the qualitative and incomplete nature of biological data, by providing for the generation, on a computer, of actual informatic ''preparations'' compatible with known results and established biochemical phenomena. Once a ''standard'' preparation is set up (much like an experimental one might be), ''experiments'' can be performed which may then guide laboratory tests or bioinformatic research. The program has already been used to study neurulation (Kerszberg and Changeux, 1998) and to examine the propagation of the morphogen activin during mesoderm formation (Kerszberg and Wolpert, 1998). But these early applications give only a hint of the actual power of ctrl-Dev, which is now being applied in particular to Insect segmentation, cochlear tufted cell morphogenesis, and the propagation of active substances in biological media.
I am in the process of preparing, with Andrei Ruckenstein (Physics Dept. and Director, BIOMAPS Center, Rutgers University) a book based on ctrl-Dev, tentatively entitled: Modeling Integrated Biological Systems / A Primer in the Unified Computation of Cellular and Developmental Processes to be published by Springer in 2003.
One particularly crucial aspect of the genotype-phenotype link concerns protein-DNA interactions. A great number of important proteins are involved, for instance, in DNA maintenance and repair; hence they control to a certain extent mutation and therefore genome evolution and phenotype variation, which in turn constitutes the raw material for selection. I have become involved, together with experimenters at Pasteur, in a complex population-theoretic and bioinformatic analysis of DNA repair systems in Plasmodium falciparum (which causes the most severe form of human malaria), in relation to the very peculiar structure of this organism's genome, its physiology, and its ecological diversity (or, sometimes, surprising absence thereof).
The central nervous system is certainly, of all organs, the one which exhibits the widest gap between genotype (whose direct expression are molecules such as ionic channels or axon guidance markers) and phenotype (i.e. behavior). This interface has been from the start the subject of my work on neural networks (Kerszberg, 1990). Most recently, I have worked on the higher cognitive functions, such the supervisory systems that enable humans to perform cognitive tasks requiring conscious effort (Dehaene, Kerszberg and Changeux, 1998). The relationship between cognitive malfunction, genetic background and stimulatory environment is at the core of this research program.
Developmental Canalization Can Enhance Species Survival. M. Kerszberg, J. Theor. Biol. 139, 287-309 (1989).
How Neurons May Compute: The Case of Insect Sexual Pheromone Discrimination. C. Linster, M. Kerszberg and C. Masson, Journal of Computational Neuroscience 1, 231-238 (1994).
A Model for Reading Morphogenetic Gradients: Autocatalysis and Competition at the Gene Level M. Kerszberg and J.-P. Changeux, Proc. Natl. Acad. Sci. (USA) 91, 5823-5827 (1994). PDF
Partners Make Patterns in Morphogenesis M. Kerszberg and J.-P. Changeux, Current Biology 4,1046-1047 (1994).
Signal-Induced Selection Among Spontaneous Oscillatory Patterns In a Model of Honeybee Olfactory Glomeruli. M. Kerszberg and C. Masson, Biological Cybernetics, 72, 487-495 (1995).
Accurate Reading of Morphogen Concentrations by Nuclear Receptors: a Formal Model of Complex Transduction Pathways. M. Kerszberg, J. Theor. Biol. 183, 95-104 (1996).PDF
Mechanisms for Positional Signalling by Morphogen Transport: a Theoretical Study. Michel Kerszberg and Lewis Wolpert, J. Theor. Biol. 191, 103-114 (1998).PDF
A Neuronal Model of a Global Workspace in Effortful Cognitive Tasks. Stanislas Dehaene, Michel Kerszberg and Jean-Pierre Changeux, Proc. Natl. Acad. Sci (USA) 95, 14529-34 (1998).PDF
The Survival of Slow Reproducers. Michel Kerszberg, J. Theor. Biol. 206, 81-89 (2000).PDF
Genes, Neurons and Codes : Remarks on Biological Communication. Michel Kerszberg, BioEssays 95, 14529-34 (2003).PDF
A current version of ctrl- Dev may be obtained upon request from the author, Michel Kerszberg . You will need to be able to compile a C++ program on your machine to be able to use this version. On a Linux machine, this should be easy since a C++ compiler comes with the system. This version can be used to examine the network of segmentation genes in Drosophila , comprised of engrailed, invected, hedgehog, patched, cubitus interruptus, and wingless. Here is a typical display where green means wingless expression, red means engrailed activity, and light to dark blue denote decreasing concentrations of extracellular wingless protein (there is an obvious anomaly developing in the segmentation pattern) :
You can visualize all genes and molecules in the system; by clicking on a cell, you can look at the genetic state of that cell, and at the local concentrations of transcription factors, signaling molecules etc. You can also look at the concentrations of diffusible molecules everywhere in the system, change parameters and consider the changes in behavior that result.
In short: you can test any idea you might have for an experiment, and see what your current theoretical understanding would predict to be the outcome.
My personal homepage (family history) : kerszberg.net
Last modification: September 17, 2004, by Michel Kerszberg