arXiv subjects

Purba Chatterjee

Publications and source records attributed to Purba Chatterjee.

6 recordsLinked to original sources

arXiv2025

Remembrance of Tasks Past in Tunable Physical Networks

Sequential learning in physical networks is hindered by catastrophic forgetting, where training a new task erases solutions to earlier ones. We show that we can significantly enhance memory of previous tasks by introducing a hard threshold in the learning rule, allowing only edges with sufficiently large training signals to be altered. Thresholding confines tuning to the spatial vicinity of inputs and outputs for each task, effectively partitioning the network into weakly overlapping functional regions. Using simulations of tunable resistor networks, we demonstrate that this strategy enables robust memory of multiple sequential tasks while reducing the number of edges and the overall tuning cost. Our results hint at constrained training as a simple, local, and scalable mechanism to overcome catastrophic forgetting in tunable matter.

cond-mat.dis-nn↗

arXiv2025

Hierarchical Loop Stabilization in Periodically Driven Elastic Networks

Network remodeling, or adaptation, in the presence of periodically driven forcings has hereto remained largely unexplored, despite the fact that a broad class of biological transport networks, e.g. animal vasculature, depends on periodic driving (pulsatility of the heart) to maintain flow. Short-term pulsatile dynamics of compliant vessels affects the long-term structures of adapting networks; however, what the correct adaptation rule is for pulsatile flows still remains an open question. Here we propose a new adaptation rule for periodically driven complex elastic networks that accounts for the effect of short-term pulsatile dynamics on the remodeling signal at long time-scales. Using this rule to adapt hierarchical elastic networks with multiple levels of looping, we show that very different network architectures are possible at steady-state depending on the driving frequency of the pulsatile source and the geometric asymmetry of the paths between the externally driven nodes of the network. Specifically resonant frequencies are shown to prioritize the stabilization of fully looped structures or higher level loops proximal to the source, whereas anti-resonant frequencies predominantly stabilize loop-less structures or lower-level loops distal to the source. Thus, this model offers a mechanism that can explain the stabilization of phenotypically diverse loopy network architectures in response to source pulsatility under physiologically relevant conditions and in the absence of other known loop stabilization mechanisms, such as random fluctuations in the load or perfusion homogenization.

nlin.AO↗

arXiv2022

Pulsatile Driving Stabilizes Loops in Elastic Flow Networks

Existing models of adaptation in biological flow networks consider their constituent vessels (e.g. veins and arteries) to be rigid, thus predicting a non physiological response when the drive (e.g. the heart) is dynamic. Here we show that incorporating pulsatile driving and properties such as fluid inertia and vessel compliance into a general adaptation framework fundamentally changes the expected structure at steady state of a minimal one-loop network. In particular, pulsatility is observed to give rise to resonances which can stabilize loops for a much broader class of metabolic cost functions than predicted by existing theories. Our work points to the need for a more realistic treatment of adaptation in biological flow networks, especially those driven by a pulsatile source, and provides insights into pathologies that emerge when such pulsatility is disrupted in human beings.

nlin.AO↗

arXiv2021

DNA Supercoiling Drives a Transition between Collective Modes of Gene Synthesis

Recent experiments showed that multiple copies of the molecular machine RNA polymerase (RNAP) can efficiently synthesize mRNA collectively in the active state of the promoter. However, environmentally-induced promoter repression results in long-distance antagonistic interactions that drastically reduce the speed of RNAPs and cause a quick arrest of mRNA synthesis. The mechanism underlying this transition between cooperative and antagonistic dynamics remains poorly understood. In this Letter, we introduce a continuum deterministic model for the translocation of RNAPs, where the speed of an RNAP is coupled to the local DNA supercoiling as well as the density of RNAPs on the gene. We assume that torsional stress experienced by individual RNAPs is exacerbated by high RNAP density on the gene and that transcription factors act as physical barriers to the diffusion of DNA supercoils. We show that this minimal model exhibits two transcription modes mediated by the torsional stress: a fluid mode when the promoter is active and a torsionally stressed mode when the promoter is repressed, in quantitative agreement with experimentally observed dynamics of co-transcribing RNAPs. Our work provides an important step towards understanding the collective dynamics of molecular machines involved in gene expression.

q-bio.SC↗

arXiv2020

A Field-Theoretic Model for Chemotaxis in Run and Tumble Particles

In this paper we develop a field-theoretic description for run and tumble chemotaxis, based on a density functional description of crystalline materials modified to capture orientational ordering. We show that this framework, with its in-built multi-particle interactions, soft-core repulsion and elasticity is ideal for describing continuum collective phases with particle resolution, but on diffusive timescales. We show that our model exhibits particle aggregation in an externally imposed constant attractant field, as is observed for phototactic or thermotactic agents. We also show that this model captures particle aggregation through self-chemotaxis, an important mechanism that aids quorum dependent cellular interactions.

cond-mat.soft↗

arXiv2018

Three-body Interactions Drive the Transition to Polar Order in a Simple Flocking Model

A large class of mesoscopic or macroscopic flocking theories are coarse-grained from microscopic models that feature binary interactions as the chief aligning mechanism. However while such theories seemingly predict the existence of polar order with just binary interactions, actomyosin motility assay experiments show that binary interactions are insufficient to obtain polar order, especially at high densities. To resolve this paradox, here we introduce a solvable one-dimensional flocking model and derive its stochastic hydrodynamics. We show that two-body interactions are insufficient to generate polar order unless the noise is non-Gaussian. We show that noisy three-body interactions in the microscopic theory allow us to capture all essential dynamical features of the flocking transition, in systems that achieve orientational order above a critical density.

cond-mat.soft↗