
Statistical Physics of Complex and Biological Systems
Our group's theoretical research activity covers scientific areas ranging from statistical mechanics to the physics of complex systems. We deal with interdisciplinary topics such as biopolymer and protein physics, liquid crystal dynamics, collective motions in self-propellant particle systems, physics of ecological systems, biological physics, and physics of complex networks. Our approach to these topics includes activities in data mining, data analysis, statistical analysis, computational modeling, and analytics.
Staff
Full Professors: Amos Maritan, Enzo Orlandini, Flavio Seno
Associate Professors: Sandro Azaele, Marco Baiesi, Fulvio Baldovin, Manlio De Domenico, Emanuele Locatelli, Samir Suweis, Antonio Trovato
Assistant Professors: Michele Allegra, Davide Bernardi, Valeria d'Andrea, Gianmaria Falasco.
Technical staff: Ramon Guevara
Post-doc
Francesco Arceri, Jean-François Derivaux, Jacopo Fadanni, Danilo Forastiere, Christian Grilletta, Johannes Nauta, Charley Presigny, Tomas Scagliarini, Vittoria Sposini, Marin Vatin
PhD students
Giacomo Barzon, Anna Braghetto, Alice Doimo, Clelia Corridori, Francesco Ferraro, Tommaso Jack Leonardi, Benedetta Mariani, Fabio Menegazzo, Jacopo Pasqualini, Emanuele Pigani, Leonardo Salicari, Davide Santolin, Marika Sartore, Kobe Simoens, Elisa Tentori, Andrea Veronese, Chiara Veronese, Giorgio Vittorio Visco, Clemens Franz Vorsmann, Meixi Yuan, Francesco Zambelli, Davide Zanchetta
Research activities
Emergence of patterns in complex systems
The universality of scaling laws is one of the most attractive features of statistical physics because a large class of patterns can be classified solely in terms of their macroscopic behaviors whose characteristics depend on only a few details, such as the size of the system and the symmetries of the problem (and possibly the way the interaction decreases), but not on more microscopic details. Laws of scale have been observed and described in many physical, chemical, biological, ecological, economic, and sociological phenomena. Our research interests include biological and ecological systems, statistical physics of interacting particles, optimal transport networks, and dynamics of complex networks with applications to environmental science and ecological sustainability. For more information visit:
Contacts: Sandro Azaele, Valeria d’Andrea, Manlio De Domenico, Amos Maritan, Samir Suweis.
Website: www.liphlab.com
- Complex Multilayer Networks Lab
Ecosystems Organization and Dynamics of Evolution
Understanding the origins, maintenance, and loss of biodiversity in ecological systems is one of the highest scientific priority goals given the alarmingly high biodiversity loss across the globe. Ecological communities show widespread patterns and interrelationships between size, species abundance and the availability of resources it can draw on. Nonequilibrium statistical mechanics is the natural candidate to try to develop a unified theory that can describe the size distribution of species, their energy use, and their relative spatial distribution. Our approach is to use variational or optimization principles, which have been very successful in physics, to try to understand and describe spatio-temporal patterns observed in natural systems (such as the architecture of ecological networks, the species-area relationship in ecosystems), so that we may also be able to predict undiscovered patterns. Wherever possible, alongside theoretical investigation and modeling results, a comparison of theory with empirical data analysis is always sought. For more information visit:
Contacts: Sandro Azaele, Davide Bernardi, Manlio De Domenico, Amos Maritan, Samir Suweis
Website: www.liphlab.com
- Complex Multilayer Networks Lab
Computational Neuroscience
Statistical physics and network science are making major contributions to describing the structure and self-organized dynamics of the human brain. Our research in computational neuroscience is developed along three directions: 1) Criticality of the brain; a recent neuroscience hypothesis is that the resting brain may be near a phase transition (in the physical statistical sense), between a sub-critical regime with low activity and a super-critical regime of high activity. We have shown that the optimal response to a changing and complex environment occurs in systems that spontaneously organize near a critical point. Using various methods, we are testing these hypotheses in rat brain activity; 2) Controllability of brain networks; we apply network control theory to neuroscience to understand whether a single brain region can control the dynamics of the entire brain through external stimulation, trying to identify brain regions that have greater control over others; 3) Whole Brain modeling; in this line of research we study the relationship between structure and function at various levels in the brain, with the goal of understanding how structural lesions (such as stroke) or other brain disorders lead to deviations from the criticality described in the first point. These lines of research are also being developed in collaboration with the Padua Neuroscience Center. For more information visit:
Contacts: Samir Suweis, Michele Allegra, Davide Bernardi, Manlio De Domenico, Valeria d’Andrea
Website: www.liphlab.com
- Complex Multilayer Networks Lab
Robustness, Adaptability and Critical Transitions in Living Systems
Understanding biological systems needs more than a simple generalization of the standard methods of statistical mechanics. In the latter, it is crucial to determine the order parameter, which characterizes the various phases of the system. This is a crucial step in obtaining the key ingredients needed to formulate an essential but complete picture of the macroscopic behavior of the system. However, what the order parameter is for a certain living system (from genomes, to human societies) is a difficult and still not well defined problem.However, there is growing evidence that the fundamental characteristic of living systems is in the architecture of their interaction networks. Our research focuses on three points: 1) Developing a unifying theoretical framework that can provide a parsimonious and general explanation for the macroscopic behavior of these systems; 2) Investigating network responses to local or global perturbations; and 3) Designing novel network architectures aimed at maximizing various objective functions in terms of system adaptability or robustness. This line has applications in various fields, including neuroscience, systems biology, and more generally complex networks. For more information visit:
Contacts: Michele Allegra, Davide Bernardi, Valeria d’Andrea, Sandro Azaele, Manlio De Domenico, Amos Maritan, Samir Suweis
Website: www.liphlab.com
- Complex Multilayer Networks Lab
Soft matter
Soft matter includes all those macromolecular systems, held together by weak intramolecular forces, whose typical energy scale is given by thermal fluctuations at room temperature. Examples are colloidal and polymeric suspensions and liquid crystals, collectively also called “complex fluids.” The group is interested in several topics:
- Heterogeneously charged colloids: we are interested in modeling heterogeneously charged colloids and proteins, thermodynamic and self-assembly properties
- Phase separation: we are interested in characterizing the thermodynamic properties of multi-component systems that exhibit separation between two liquid phases, a phenomenon of considerable importance in cells
- Diffusion in heterogeneous and glassy systems: we study transport phenomena in heterogeneous systems, with anomalous diffusion properties or characterized by kinetic traps and metastable states (such as glasses), with attention to a microscopic description.
- Active microrheology: complex fluids often have viscoelastic properties, that is, they behave partially as viscous fluids and partially as elastic solids. In active microreology, the dynamics of a microscopic probe are modeled and interpreted to reveal the properties of the complex fluid in which it is immersed.
Our approach to the various topics includes numerical simulations (Monte Carlo and molecular dynamics)
Contacts: Marco Baiesi, Fulvio Baldovin, Gianmaria Falasco, Emanuele Locatelli, Enzo Orlandini, Flavio Seno, Antonio Trovato
Sito web: BSSP Group
Dynamics and thermodynamics of polymeric systems
Polymers are macromolecules, composed of repeated elements: they are ubiquitous, from the basic components of life such as proteins, DNA and RNA to commonly used materials such as plastics, rubbers and textile fibers. Polymers can take on a number of spatial configurations that grow exponentially with their length: a theoretical description then becomes a problem of statistical mechanics, and the pioneering work of De Gennes showed how polymers have properties that do not depend strictly on their chemical details but are, on the contrary, universal. The group is interested in several topics:
- Polymer dynamics: we are interested in transport properties in complex systems under confinement, such as in porous materials or in channels of varying amplitude.
- Polymer thermodynamics: we study phase transition phenomena in polymers and biopolymers, including the collapse transition from extended to globular phase in chromatin models or DNA denaturation.
- Absorption properties: we are interested in characterizing the phase transition, related to material adsorption in suspension, as the topology, architecture (i.e., how we organize monomers) and geometry of the problem change.
- Coarse-grained polymer models: to study the material properties of a polymer system, it is often necessary to remove unimportant details to make the problem computationally tractable. We apply these techniques to polymers with complex topology and/or architecture relevant to the study of biosystems.
Our approach to various topics includes numerical simulations (Monte Carlo and molecular dynamics) and analytical approaches such as field theories.
Contatti: Marco Baiesi, Fulvio Baldovin, Gianmaria Falasco, Emanuele Locatelli, Enzo Orlandini, Flavio Seno, Antonio Trovato
Website: BSSP Group
Topological properties in polymeric systems
Topology is a branch of mathematics concerned with those properties of curves that do not depend on conformation but on global properties, which do not change except by affecting the integrity of the curves themselves (e.g., by cutting and stitching). For physical systems, the presence of topological properties induces peculiar material properties. We are interested in understanding how the properties of polymeric systems with well-defined topology (e.g., concatenated rings) depend on factors such as their stiffness, confinement, or solvent quality; in addition, we are interested in understanding how topological properties affect macroscopic properties (such as viscoelasticity) of the system. Our group has developed methods to identify knotted regions in polymers: this allows us to identify and characterize knots in physical systems of interest. Finally, topology is also relevant to biological systems: open problems include understanding how DNA in cells is knotted by topoisomerases.
Contacts: Marco Baiesi, Emanuele Locatelli, Enzo Orlandini
Website: BSSP Group
Proteins and biopolymers
Biopolymers in general, and proteins in particular, are essential for living organisms. The problem of determining the native state of a protein is extremely complex, given the large number of degrees of freedom in the system. Non-native protein conformations are also of extreme interest, as they can trigger pathological protein aggregation, the cause of various degenerative diseases. Our mechanistic-statistical approach is key to understanding the unifying aspects that emerge in protein physics and their relationship to ordinary polymers. Some specific topics include understanding the origin of native protein structures by means of geometry- and symmetry-based considerations; developing algorithms, based on both sequence and structure, to predict various protein features; and studying the major mechanisms governing protein folding and aggregation. In particular, we are interested in the folding mechanisms of proteins characterized by topological complexity in the native structure and the presence/characterization of intermediate states with topological properties different from the native ones. Their formation during protein synthesis could indeed play a relevant role in deteriorating cellular proteostasis. Recently, we have also recently applied machine learning methods to study the properties of amino acids, from which emerged an effective hydrophobicity table that reveals aspects of how Nature employs different types of amino acids to determine secondary structures such as alpha-helices and beta-sheets in proteins.
Contacts: Marco Baiesi, Amos Maritan, Enzo Orlandini, Flavio Seno, Antonio Trovato
Website: BSSP Group
Rheological and dynamic properties of passive and active liquid crystals
Liquid crystals are fluids typically formed by long, thin molecules that, under certain environmental conditions, tend to align to form ordered states of varying complexity such as nematic, chiral and smectic phase. They are important examples of structured fluids that respond to external stresses either as elastic materials or as viscous fluids. Liquid crystals are also important from an application point of view because of their well-known use in digital devices, the operation of which relies on the system's orientational response (switching dynamics) with respect to the switching on/off of external electric fields. Moreover, when suitably modified with terms that cannot be attributed to a free energy, they are a good macroscopic description of active fluids such as actin solutions in the presence of molecular motors. We are concerned with the study, by means of numerical simulations (Lattice Boltzmann methods) and mean-field theories, of the rheological and dynamic properties of these systems when subjected to external electric or velocity fields or as activity changes.
Contatti: Enzo Orlandini
Sito web: BSSP Group
Dynamics of self-propellant systems
“Active” self-propellant systems are composed of units that draw energy from their surroundings (or from an internal store) and burn it to move autonomously. The individual self-propelling process keeps the system out of thermodynamic equilibrium making them extremely different from their “passive” versions as well as other forms of non-equilibrium. Examples of these systems are ubiquitous, from the micro-scale, such as the cellular cytoskeleton and bacterial suspensions, to the macro-scale, such as worm agglomerates and bird flocks. In particular, those proposed are examples of self-propelled and interacting many-body systems that exhibit spontaneous collective motion phenomena. Our research focuses on simplified models and the study of their mechanistic-statistical properties. The group is interested in several topics
Segregation of active colloids: in particular, those that result from competing spatial confinement, effective interactions and communication mechanisms between individuals.
Properties of self-propellant polymers: we study the statistical properties of self-propellant polymers in dependence on their polymeric properties (length, stiffness and architecture) and external conditions (confinement and density).
Biophysical models: we are interested in applying active polymer models for biophysical systems, particularly to understand their minimal ingredients
We use stochastic simulations, analytical techniques (Smoluchowski-type equations). The models considered describe individuals either as asymmetric Brownian particles with directional internal force, point objects with a certain position, direction of motion, and constant velocity, or minimal models of self-propellant polymers.
Contatti: Fulvio Baldovin, Emanuele Locatelli, Enzo Orlandini
Sito web: BSSP Group
Statistical mechanics out of equilibrium
One of the current challenges is to develop a general theory for out-of-equilibrium systems. These systems, subjected to nonconservative forces, maintain fluxes of heat or matter and dissipate free energy. These are often small systems (e.g., living cells), where non-Gaussian fluctuations and nonlinear effects play a crucial role in determining dynamics and thermodynamics. Major research topics include response to external perturbations, generalization of concepts such as energy equipartitioning, stability of dissipative patterns with respect to thermal noise, efficiency in energy transduction, and thermodynamic characterization of out-of-equilibrium phase transitions. Free energy dissipation in individual microscopic processes is of particular importance, both in biological systems and in technological applications. Therefore, we develop inference methods capable of estimating the magnitude of such dissipation.
Contacts: Marco Baiesi, Fulvio Baldovin, Gianmaria Falasco, Amos Maritan
Website: BSSP Group
Statistical Physics and Machine Learning
Nowadays there is a growing interest in solving complex problems using machine learning techniques, known as machine learning. This is leading us to evaluate new ideas and tools from the community studying machine learning, with the goal of applying these methods in contexts such as complex systems, neural networks, or polymeric phases. One goal is to understand how machine learning works in distinguishing polymeric, epigenetic or brain phases. A general plan is to apply machine learning techniques to infer emergent patterns in complex ecological, biological, social, and geophysical systems. Finally, we would like to understand whether and in what sense neural networks are critical and how their performance can be understood through information theory and statistical physics.
Contatti: Amos Maritan, Enzo Orlandini, Marco Baiesi, Samir Suweis, Manlio De Domenico, Valeria d’Andrea
Statistical Physics of Complex Networks
A wide variety of complex systems, from cells to societies, are characterized by a nontrivial structure: a complex network. Networks are ubiquitous and are successfully used to describe the interactions and relationships among physical/biological units (such as proteins, neurons, computers, power plants, etc.), active matter (such as molecular motors, birds and other living systems) and, in some cases, even individuals. Surprisingly, empirical complex networks exhibit mesoscale organization, heterogeneous connectivity patterns, and long-range correlations, which are responsible for many fascinating collective phenomena and complex functions, but which also make them difficult to analyze with standard analytical techniques.
Our research interests lie between theoretical and applied aspects of network science: we use multilayer models (multilayer), network geometry (network geometry), and information dynamics to study structure, dynamics, and functions (e.g., robustness to perturbations) of (i) empirical biological networks at the cellular scale, (ii) biological and artificial neuronal systems, (iii) social and socio-technical networks, and (iv) infrastructure networks (from transportation to communication). Applications range from precision medicine (network medicine) to epidemic spreading and computational social sciences.
Contacts: Manlio De Domenico, Valeria d’Andrea, Michele Allegra, Davide Bernardi, Sandro Azaele, Amos Maritan, Samir Suweis.
Website: www.liphlab.com
- Complex Multilayer Networks Lab