Research Interests
The number of known DNA and protein sequences is growing extremely
fast. Nevertheless, a number of fundamental questions are still open
(gene regulation in eukaryots; origin and role of introns and repeats;
long-range correlations; modular structure of genes and proteins;
protein folding). In order to analyze these questions, suitable statistical
and information-theoretical concepts are necessary. For the immediate
analysis of new sequences, a number of algorithms is available. The
main point of our project lies, therefore, in the detection and analysis
of general genome structures (periodicities in DNA- und protein sequences,
long-range correlations, redundance by repetitive sequences, modularity
of genomes).
Building on methods from statistical physics (correlation functions,
mutual information, entropies), statistical dependences in sequences
are analyzed. There are strong indications that interesting general
aspects of the evolution of genomes can be made accessible from such
statistical examinations. It has been shown, for instance, that certain
statistical properties of the mutual information function of DNA sequences
are universal, i.e. they are the same for different taxonomic classes
(vertebrates, invertebrates, plants). Periodicities of 10-11 basepairs
in complete genomes point to the supercoiled state of the DNA: negative
supercoiling in eubacteria and positive supercoiling in archaea.
High density DNA-arrays ("DNA Chips")
allow measurements of gene expression levels for a large number of
genes simultaneously.
In this way thousands of mRNA concentrations can be analyzed
in parallel, potentially revealing complex gene regulatory networks.
In close collaboration
with the MPI
for Molecular Genetics (H. Lehrach) and cell biologists
at the Charité (C. Sers, R. Sch?r) we assess the data reliability
(image analysis, calibration, reproducibility), identify co-regulated
genes by cluster analysis, and detect transcription factor binding
sites in clusters of co-regulated genes. It is our aim to incorporate
the resulting information into network models of signaling cascades
such as the Ras pathway.
Periodic and complex oscillations
play a central role in biological systems. Examples include physiological
rhythms (heartbeat, respiration,
blood-pressure waves), neural oscillations and sound generation
for acustical communication. The conceptual framework for the theoretical
description of generation and interaction of rhythms is provided
by
the theory of dynamical systems. The general theory has led to
a detailed understanding of attractors, bifurcations, phase-response-curves
etc.,
whereas quantitative models of physiological systems are still
rare. The theory of nonlinear dynamical systems allows us to view complex
biomedical systems from a new perspective. Irregularities can
partially
be interpreted as a signature of deterministic chaos, and instabilities
of parameters may be understood as bifurcations of the underlying
dynamical systems. The new concepts have far-reaching consequences
for signal
analysis (attractor analysis in addition to the traditional spectral
analysis) as well as for building adequate models.
The immediate goal of this project is to develop and test nonlinear
models for two special systems (vocal fold oscillations, atrial
fibrillation) on the basis of experimental data. The examination
of vocal fold oscillations
is performed in close collaboration with the National
Center for Voice and Speech of the USA and the university hospitals
Charité,
Berlin-Steglitz and Erlangen. Together with the experimental
biologists at the Humboldt
University and T. Fitch (Harvard) nonlinear phenomena (subharmonies,
biphonation, chaos) are studied in animal vocalization. Models
developed to described human phonation are adapted to additional
features found in non-human vocalizations such as vocal membranes
and air sacs. The model of the atrium is developed together with
A. Panfilov
in Utrecht.
List of Publications
List of Publications (PDF) |
Prof Dr. Hans-Peter Herzel
