Sommaire:
Stochastic processes have been in the last decades very efficient for modelling biological systems, essentially because of the natural randomness that governs them. Genetic network regulation, molecular transport inside cells, and metabolic processes all share high randomness and variability, taking place at finite temperature. Life itself is an emergent process that came out of randomness. This is why complex systems physics have a central role for understanding living systems. In this seminar I will describe three biological processes where a physics approach has helped in the understanding of the phenomenon and to interesting generalizations.
The first one concerns avalanches of fractures. Many glassy and amorphous materials, such as martensites, exhibit characteristic behaviours during constraint-induced fractures. These fractures manifest as avalanche processes, with statistics often following a power-law distribution, reminiscent of collective behaviour and self-organized criticality. Similar avalanches of fractures occur in living systems, resembling a glassy network with a frozen structure, notably observed in the actin cytoskeleton (CSK). The CSK comprises microfilaments organized into higher-order structures via dynamic assembly-disassembly mechanisms with cross-linkers. Experiments indicate that cells respond to external constraints with a cascade of random and abrupt ruptures of their CSK. These behaviours are analysed in experimental data on CD34+ cells from healthy and leukemic bone marrows, revealing log-normal distribution of rupture events. To interpret this behaviour, a minimal (1D) stochastic model is proposed, explaining energy released along rupture events as a sum of a multiplicative cascade process. The model distinguishes between brittle failures, corresponding to irreversible ruptures, and ductile failures, resulting from dynamic cross-linker unbindings. Brittle failures are relatively more prominent in leukemic cells, suggesting greater fragility and different CSK architecture. Further analysis explores the distribution of a sum of correlated random variables from a multiplicative process, highlighting regions of log-normal and power-law distributions in a phase diagram.
The second one is a chemical reaction network, the dual phosphorylation/dephosphorylation cycle, which allows the cell to switch between active and inactive states. I introduce the Random Walk Approximation (RWA), a new method to approximate the stationary solution of master equations describing stochastic processes taking place on graphs. Our approach is suitable for non-linear master equations with conserved entities, common also in other biological systems like gene regulatory networks. The RWA allows having a simple analytical, even though approximated, form of the solution, which is global and easier to deal with than the standard System Size Expansion (SSE). Here, I give some theoretically sufficient conditions for the validity of the RWA and show its interest in biological predictions.
The third one is about the concept of Synthetic Lethality to treat blood cancer (AML). Differential expression analysis of RNA-Seq data provide primary molecular markers that are propagated in protein-protein interaction networks, to identify secondary not altered candidate genes of Synthetic Lethality. I present the physical foundation of this method, which enables to rank the gene based on their “diffusive power”. The identified markers are first reduced by the ranking given by the diffusion and by the centrality of the gene inside the network and then DNA Damage Response (DDR)-related genes for which a drug inhibitor is available are selected. The resulting drugs will be screened ex-vivo on primary leukemic cells. Results showing the efficacy of the method in selecting potentially successful therapy on tumoral cell lines will be presented.
Finally, the importance of physical modelling in biological systems is mainly embodied by the creation of phase diagrams which create possible unexplored scenario to be experimentally tested.
Pour plus d'informations, merci de contacter Palmeri J.