|
|
A poster session will be organised during the workshop
From DNA sequence to chromatin organization: the fundamental role of genomic long-range correlations
Phylogenies and Distances
Gene regulation, interactions, and gene expression dynamics
Statistical modelling and analysis of biological networks.
Detecting functionality in biological networks is a major goal of systems biology. Such networks consist of functional units in an effectively random background, so we need statistical models and algorithms to discriminate both parts. In this talk I discuss a statistical theory of networks, using the evolutionary dynamics of nodes and links to distinguish functional from random parts. Three particular cases are considered: clusters within a network, repetitive network motifs, and cross-species correlations between networks, with examples from protein interaction networks, transcriptional regulation networks, and co-expression networks.
Computational search of mixed miRNA-Transcription Factors regulatory circuits.
The expression of coding genes is tightly regulated both at the transcriptional level (by Transcription Factors) and at the post-transcriptional level (mainly by microRNAs). While lot of results by now exist separately on transcription factor and microRNA-related regulatory networks, comparable information is lacking to explicitly connect them. In this talk I will first describe a genome-wide search of mixed regulatory circuits in which a master transcription factor regulates a microRNA and together with it a set of joint target coding genes. Then, in the second part of the talk, I will discuss the main properties of this type of circuits and describe a few biologically relevant examples.
DNA bending stress in molecular dynamics simulations.
Many biological functions directed by DNA bending flexibility reflect dynamical hinge properties between specific base pair components. Recent experiments on minicircles [1] suggest that this DNA propensity is much higher than expected from the canonical persistence length of 50 nm. Subsequent molecular dynamics simulation of these minicircles [2] found base-pair kink motifs referred to as type I (adjacent base pair unstacking without base pair disruption) and type II (strong unstacking associated with base pair disruption) provide ways to relax DNA elastic energy. We impose a controlled bending deformation of short DNA oligonucleotides during molecular dynamics [3] to estimate free energy costs and dynamical bending response of DNA as a function of sequence. We summarize the results in terms of base pair kink motifs and global DNA bending. Local DNA stiffness constants depend significantly on the nature of such motifs, and can reduce canonical DNA values by as much as an order of magnitude in the case of type II kink formation.
1.T.E. Cloutier, J. Widom (2004) Mol. Cell., 14: 355-62. 2.F. Lankas, R. Lavery, J.D. Maddocks (2006) Structure, 14: 1527-34. 3.J. Curuksu, K. Zakrzewska, M. Zacharias (2008) Nucleic Acids Res., 36: 2268-83.
p-Adic Approach to the Genetic Code and the Genome
Abstract of talk: We present the foundations of p-adic approach to genomics. Nucleotides, codons, DNA and RNA sequences, amino acids and proteins are considered as p-adic information systems. Each of these systems has its characteristic prime number which generate the related information space. It is shown that degeneration of the genetic code is a p-adic phenomenon. We also propose evolution of the genetic code assuming that primitive code started by single nucleotides and the first four amino acids (Gly, Ala, Asp and Val).
References: arXiv: q-bio.GN/0607018, arXiv:07070764[q-bio.GN], arXiv:0707.3043[q-bio.OT] .
Evolution and Turing machines.
What are the Turing machines? How can they be related to the coding Vs non-coding parts in DNA? I explain these concepts and introduce a toy model to study the darwinian evolution of "organisms" that are selected according to the principle of the "survival of the fittest". I emphasize the importance of the so called non-coding states in the evolution process.
A new approach towards deciphering the protein code: The protein assembly model.
Proteins are responsible for all activities in a living organism. Each protein chain is made of a specific sequence of amino acids, called the primary sequence. The 20 existing amino acids constitute the building unit of proteins. In the early seventies, Anfinsen showed that the primary sequence of a protein contained all the necessary information for a protein to fold into its functional 3D shape (3D structure). Yet, the mechanism by which a protein acquires its 3D structure, the so-called protein folding is far from understood. What remains particularly mysterious is how the information for the 3D structure is encoded within the amino acids of the primary sequence. Thus, understanding the protein code remains an extremely challenging issue.
The vast majority of proteins are oligomers containing two -or more- copies of the protein chain. The specificity of protein oligomers is their capacity to form interfaces between different chains, in order to assemble. To be made, protein interfaces use a lock and key mechanism. The areas forming the interfaces have a geometrical constraint plus a well defined set of amino acid interactions. Hence, the amino acids involved in protein interfaces must share some features to be able to fulfill these two constraints.
To address the problem of the protein code in a simplified manner, we propose to focus on understanding the sequence-structure relationship in proteins interfaces.
Joint work with G. Feverati and P. Sorba.
Calibration of a molecular thermometer to predict Topt from proteome composition.
In my talk I will present the effect of temperature T on microbial specific growth rate, µ : asymmetry of the response curve and the correlation between the three cardinal temperatures, Tmin, Topt and Tmax. The connection with the optimal temperature for enzymes, Tm, will be discussed. The between proteome variability in amino-acid content will be analyzed from a dataset of 739 proteomes (\textit{Gene} (2006) \textbf{385}:128-136) focusing on the biological interpretation of the two main factors (\textit{viz.} G+C content and temperature) of variability. Three previous attempts to define a molecular thermomether to predict Topt from amino-acid frequencies will be presented and a new one introduced. The last one will be applied to the recently inferred proteome composition of the Last Universal Common Ancestor, LUCA (\textit{Nature} (2008) \textbf{000}:000-000).
A Symmetry Breaking Model for X Chromosome Inactivation.
In humans, female cells silence one of their two X chromosomes, which is randomly selected, to equalize X products with respect to males (having just one X). Such a process, named X Chromosome Inactivation (XCI), is crucial to survival and is related to serious genetic diseases. The mechanism, though, whereby cells count their X's and randomly choose the one to inactivate is one of the most mysterious aspects in X biology.
Starting from recent experimental discoveries, we proposed a Statistical Mechanics model of XCI where a molecular complex, a 'blocking factor' (BF), can protect from inactivation the X chromosome which it binds to. The BF is present in a single copy in the nucleus so just one X per cell, randomly selected, can be protected, as the second X is inactivated by default under the action of the Xist gene. A crucial step in the model was to explain how the molecular complex is self-assembled and why only one is formed out of many diffusible molecules. We showed this is the result of a thermodynamics phase transition which spontaneously breaks the symmetry between the X's.
As more recent experiments have indeed discovered complexing and binding molecules regulating XCI, the mechanism that directs the two chromosomes to opposite fates appears to be clarified. In a broader perspective, at least 10% of our genes has a behavior similar to the X's, i.e., out of two alleles one is randomly selected and inactivated, with important and poorly understood examples ranging from the immune system to our olfactory apparatus. The new stochastic regulatory mechanism we propose can be a key to those cases as well.
Experimental and theoretical studies of sequence effects on the fluctuation and melting of short DNA molecules.
Understanding the melting of short DNA sequences probes DNA at the scale of the genetic code and raises questions which are very different from those posed by very long sequences, which have been extensively studied. We investigate this problem by combining experiments and theory. A new experimental method allows us to make a mapping of the opening of the guanines along the sequence as a function of temperature. The results indicate that non-local effects may be important in DNA because an AT-rich region is able to influence the opening of a base pair which is about 10 base pairs away. An earlier mesoscopic model of DNA is modified to correctly describe the time scales associated to the opening of individual base pairs well below melting, and to properly take into account the sequence. Using this model to analyze some characteristic sequences for which detailed experimental data on the melting is available [Montrichok et al. 2003 Europhys. Lett. 62 452], we show that we have to introduce non-local effects of AT-rich regions to get acceptable results. This brings a second indication that the influence of these highly fluctuating regions of DNA on their neighborhood can extend to some distance.
Joint work with Santiago Cuesta-López and Dimitar Angelov.
RNA folding and matrix field theory.
After reviewing some basic biological and chemical properties of RNA, such as secondary structures or pseudo-knots, I will show how the problem of RNA folding can be mapped onto an NxN matrix field theory. This mapping allows for a natural classification of RNA structures in terms of their topological genus: secondary structures are identified as planar structures, whereas pseudo-knots are structures with higher genus. Using this classification, I will show how RNA structures can be predicted either by recursion relations or by Monte Carlo simulations.
Analysis of feedback circuits in discrete models of regulatory networks.
The important role of feedback circuits in the dynamical behaviour of biological systems is well-known. We focus here on the analysis of such motifs in regulatory networks, modelled within a discrete framework. One first natural step is to study the dynamics of isolated circuits. Then, we concentrate on "Thomas' rules", which make a link between dynamical properties of the system, and structural properties of its regulatory graph: the occurence of multistability (differentiation) or attractive cycles (homeostasis) in the dynamics implies the presence of circuits in the
corresponding regulatory graphs. Theses rules have been enounced by mathematicians as conjectures, and then proved in different formalisms. We present here these rules enounced in a Boolean formalism, and some generalisations.
Joint works with Claudine Chaouiya, Denis Thieffry (TAGC, Inserm Marseille) and Paul Ruet (Institut de Mathématiques de Luminy, Marseille).
Sequence-encoded energy barriers determine highly-ordered chromatin organization in yeast genes.
Chromatin is composed of regularly spaced nucleosomes in which ~146 base pairs of genomic DNA is wrapped around a complex of histone proteins. Whether or not a defined genomic region is occupied by a nucleosome is essential for the binding of this region to regulatory proteins. Several recent in vivo studies of nucleosome positioning suggest that DNA sequence can be favorable or unfavorable to nucleosome formation and thus can play a significant role in organizing nucleosomal arrays. In particular, genes are bordered by nucleosome-free regions (NFR) at both extremities (5’ and 3’) and the 5’ NFR forms a barrier against which the first nucleosome aligns (1). We analyzed genome-wide nucleosome positioning data and observed that chromatin exhibits a higher level of organization than previously described. If we order the genes by the distance l that separates the first (5’) and the last (3’) nucleosomes, the resulting 2-D map reveals a strikingly organized nucleosome distribution presenting series of periodic patterns alternating with non-periodic patterns. Periodic patterns correspond to genes with regular distributions of nucleosomes between the 5’ and 3’ NFRs resulting in a “crystal-like” ladder. Non-periodic patterns correspond to genes that can form ladders with n or n+1 nucleosomes, leading to a bi-stable chromatin structure. We observed that occurrences of l values corresponding to the bi-stable chromatin patterns are strongly under-represented suggesting that the properties associated with these patterns would not be appropriate for gene expression regulation. Using an elastic model based on sequence dependent physical properties of the DNA double helix (2,3), we compare the computed and the experimental nucleosome occupancy profiles and demonstrate that nucleosome assembly between the 5’ and 3’ NFR reproduces by a “parking-like” process the observed alternation of regular and bi-stable patterns. Remarkably, the genes presenting a bi-stable chromatin pattern are enriched (i) in genes presenting a high degree of transcriptional plasticity, i.e. a large capacity capacity to modulate gene expression upon changing conditions, and (ii) in genes whose nucleosome position can be modified by a chromatin remodelling complex. These results shed a new light on the role of chromatin structure and dynamics on gene expression regulation.
Joint work with Claude Thermes1, Cédric Vaillant2, Guillaume Chevereau2, Leonor Palmeira2, Yves d'Aubenton-Carafa1, Benjamin Audit2 and Alain Arneodo2
1 Centre de Génétique Moléculaire (CNRS), Allée de la Terrasse, 91198 Gif-sur-Yvette, France
2 Laboratoire Joliot Curie et Laboratoire de Physique, Ecole Normale Supérieure de Lyon, CNRS, 69364 Lyon, France
Geometric constraints on virus architecture and their implications for virus assembly and viral evolution.
It has long been recognised that icosahedral symmetry plays a crucial role in virus architecture. For example Caspar and Klug used this fact in their seminal work in 1962 to predict the structural organisation of the viral protein containers, called viral capsids that encapsulate and hence provide protection for the viral genomic material. We introduce here an extension of this symmetry principle, based on affine extensions of the icosahedral group, and show that in this way also the structures and radial extensions of all material boundaries in simple RNA viruses can be predicted. In particular, our results show that there are collective geometric constraints on the structure of the protein container of a virus and that of its packaged RNA. As an application of our theory, its implications for virus assembly and viral evolution are discussed.