arXiv subjects

Arvind Murugan

Publications and source records attributed to Arvind Murugan.

At least 19 recordsLinked to original sources

arXiv2026

Decoding molecular distributional codes through collective instabilities

Biological information is often encoded in molecular variants that differ in just a few chemical traits, such as the number of phosphorylated sites or ubiquitin chain length, rather than in arbitrarily distinct species. These molecular distributions carry information about cellular state, yet reading them with conventional molecular circuits requires prohibitively many distinct sensors. In contrast, we show that collective physical instabilities can naturally integrate the information encoded in such distributions. Using an information-theoretic matching condition between an encoded distribution and a physical readout, we derive a geometric condition that any good decoder must satisfy, and establish that phase separation, percolation, and membrane curvature instabilities all approach it for biologically natural distributions while simple mass-action binding does not. Using mean-field theory and lattice Monte Carlo simulations, we find that phase separation reads the shape of a distribution beyond its mean, robustly capturing its variance and, more weakly, its skewness, whereas mass-action binding detects only the mean. Near phase boundaries the readout captures nearly all the information present in the molecular population. Finite valency, through the threshold for network formation, adds discriminatory power invisible to mean-field theory. These results suggest that cells can exploit collective physical instabilities as natural, compact, yet near-optimal sensors for decoding molecular distributional codes.

physics.bio-ph↗

arXiv2026

Combinatorial decision-making driven by multicomponent surface condensates

Living organisms rely on molecular networks, such as gene circuits and signaling pathways, for information processing and robust decision-making in crowded, noisy environments. Recent advances show that interacting biomolecules self-organize by phase transitions into coexisting spatial compartments called condensates, often on cellular surfaces such as chromatin and membranes. In this paper, we demonstrate that multicomponent fluids can be designed to recruit distinct condensates to surfaces with differing compositions, performing a form of surface classification by condensation. We draw an analogy to multidimensional classification in machine learning and explore how hidden species, analogous to hidden nodes, expand the expressivity and capacity of these interacting ensembles to facilitate complex decision boundaries. By simply changing levels of individual species, we find that the same molecular repertoire can be reprogrammed to solve new tasks. Together, our findings suggest that the physical processes underlying biomolecular condensates can encode and drive adaptive information processing beyond compartmentalization.

physics.bio-ph↗

arXiv2026

Conditioning as a route to stereotyped behavior in growing populations

Biological systems perform complex multi-step processes in a reproducible way despite underlying stochasticity. The standard explanation is micromanagement by molecular machinery that recognizes and corrects specific errors. Here we study conditioning, a qualitatively different strategy in which attempts failing a coarse criterion are destroyed and do not leave a physical record. The surviving, i.e., conditioned, ensemble is narrower and therefore more ordered. We model conditioning through stochastic resets in a ''socks-before-shoes'' model of a growing population, where $n$ actions must be completed in any order to replicate and any replication attempt not finished by a threshold time is discarded. We find that resets impose hierarchical temporal ordering of the $n$ actions without microscopic control over which action happens when. When disorder carries a sufficient time penalty, this ordering is free: the fastest-growing population is automatically the most ordered, with no direct selection for order required. Save points, at which verified progress is preserved across resets, allow conditioning to scale to complex multi-step processes. Conditioning provides a minimal route to reliable behavior, requiring only a clock rather than molecular machinery that recognizes specific errors. For the right class of processes, it pays for itself.

q-bio.PE↗

arXiv2025

Theoretical limits for sensing through phase separation

Biomolecular condensates form on timescales of seconds in cells upon environmental or compositional changes. Condensate formation is thus argued to act as a mechanism for sensing such changes and quickly initiating downstream processes, such as forming stress granules in response to heat stress and amplifying cGAS enzymatic activity upon detection of cytosolic DNA. Here, we study a dynamical model of droplet nucleation and growth to demonstrate how phase separation allows cells to discriminate small concentration differences on finite, biologically relevant timescales. We propose optimal sensing protocols, which use the sharp onset of phase separation. We show how, given experimentally measured rates, cells can achieve rapid and robust sensing of concentration differences of 1% on a timescale of minutes, offering an alternative to classical biochemical mechanisms.

q-bio.SC↗

arXiv2025

Learning to generalize in evolution through annealed population heterogeneity

Evolutionary systems must learn to generalize, often extrapolating from a limited set of selective conditions to anticipate future environmental changes. The mechanisms enabling such generalization remain poorly understood, despite their importance to predict ecological robustness, drug resistance, or design future-proof vaccination strategies. Here, we demonstrate that annealed population heterogeneity, wherein distinct individuals in the population experience different instances of a complex environment over time, can act as a form of implicit regularization and facilitate evolutionary generalization. Mathematically, annealed heterogeneity introduces a variance-weighted demographic noise term that penalizes across-environment fitness variance and effectively rescales the population size, thereby biasing evolution toward generalist solutions. This process is indeed analogous to a variant of the mini-batching strategy employed in stochastic gradient descent, where an effective multiplicative noise produces an inductive bias by triggering noise-induced transitions. Through numerical simulations and theoretical analysis we discuss the conditions under which variation in how individuals experience environmental selection can naturally promote evolutionary strategies that generalize across environments and anticipate novel challenges.

q-bio.PE↗

arXiv2025

Suppression of errors in collectively coded information

Modern life largely transmits genetic information from mother to daughter through the duplication of single physically intact molecules that encode information. However, copying an extended molecule requires complex copying machinery and high fidelity that scales with the genome size to avoid the error catastrophe. Here, we explore these fidelity requirements in an alternative architecture, the virtual circular genome, in which no one physical molecule encodes the full genetic information. Instead, information is encoded and transmitted in a collective of overlapping and interacting segments. Using a model experimental system of a complex mixture of DNA oligomers that can partly anneal and extend off each other, we find that mutant oligomers are suppressed relative to a model without collective encoding. Through simulations and theory, we show that this suppression of mutants can be explained by competition for productive binding partners. As a consequence, information can be propagated robustly in a virtual circular genome even at mutation rates expected under prebiotic conditions.

q-bio.PE↗

arXiv2025

Could Living Cells Use Phase Transitions to Process Information?

To maintain homeostasis, living cells process information with networks of interacting molecules. Traditional models for cellular information processing have focused on networks of chemical reactions between molecules. Here, we describe how networks of physical interactions could contribute to the processing of information inside cells. In particular, we focus on the impact of biomolecular condensation, a structural phase transition found in cells. Biomolecular condensation has recently been implicated in diverse cellular processes. Some of these are essentially computational, including classification and control tasks. We place these findings in the broader context of physical computing, an emerging framework for describing how the native dynamics of nonlinear physical systems can be leveraged to perform complex computations. The synthesis of these ideas raises questions about expressivity (the range of problems that cellular phase transitions might be able to solve) and learning (how these systems could adapt and evolve to solve different problems). This emerging area of research presents diverse opportunities across molecular biophysics, soft matter, and physical computing.

physics.bio-ph↗

arXiv2025

Simple biological controllers drive the evolution of soft modes

Biological systems, with many interacting components, face high-dimensional environmental fluctuations, ranging from diverse nutrient deprivations to toxins, drugs, and physical stresses. Yet, many biological control mechanisms are `simple' -- they restore homeostasis through low-dimensional representations of the system's high-dimensional state. How do low-dimensional controllers maintain homeostasis in high-dimensional systems? We develop an analytically tractable model of integral feedback for complex systems in fluctuating environments. We find that selection for homeostasis leads to the emergence of a soft mode that provides the dimensionality reduction required for the functioning of simple controllers. Our theory predicts that simple controllers that buffer environmental perturbations (e.g., stress response pathways) will also buffer mutational perturbation, an equivalence we test using experimental data across ~5000 strains in the yeast knockout collection. We also predict, counterintuitively, that knocking out a simple controller will \emph{decrease} the dimensionality of the response to environmental change; we outline transcriptomics tests to validate this. Our work suggests an evolutionary origin of soft modes whose function is for dimensionality reduction in and of itself rather than direct function like allostery, with implications ranging from cryptic genetic variation to global epistasis.

physics.bio-ph↗

arXiv2025

Ecosystems as adaptive living circuits

Unlike many physical nonequilibrium systems, in biological systems, the coupling to external energy sources is not a fixed parameter but adaptively controlled by the system itself. We do not have theoretical frameworks that allow for such adaptability. As a result, we cannot understand emergent behavior in living systems where structure formation and non-equilibrium drive coevolve. Here, using ecosystems as a model of adaptive systems, we develop a framework of living circuits whose architecture changes adaptively with the energy dissipated in each circuit edge. We find that unlike traditional nonequilibrium systems, living circuits exhibit a phase transition from equilibrium death to a nonequilibrium dissipative state beyond a critical driving potential. This transition emerges through a feedback mechanism that saves the weakest edges by routing dissipation through them, even though the adaptive rule locally rewards the strongest dissipating edges. Despite lacking any global optimization principle, living circuits achieve near-maximal dissipation, with higher drive promoting more complex circuits. Our work establishes ecosystems as paradigmatic examples of living circuits whose structure and dissipation are tuned through local adaptive rules.

q-bio.PE↗

arXiv2025

Temporal Contrastive Learning through implicit non-equilibrium memory

The backpropagation method has enabled transformative uses of neural networks. Alternatively, for energy-based models, local learning methods involving only nearby neurons offer benefits in terms of decentralized training, and allow for the possibility of learning in computationally-constrained substrates. One class of local learning methods constrasts the desired, clamped behavior with spontaneous, free behavior. However, directly contrasting free and clamped behaviors requires explicit memory. Here, we introduce `Temporal Contrastive Learning', an approach that uses integral feedback in each learning degree of freedom to provide a simple form of implicit non-equilibrium memory. During training, free and clamped behaviors are shown in a sawtooth-like protocol over time. When combined with integral feedback dynamics, these alternating temporal protocols generate an implicit memory necessary for comparing free and clamped behaviors, broadening the range of physical and biological systems capable of contrastive learning. Finally, we show that non-equilibrium dissipation improves learning quality and determine a Landauer-like energy cost of contrastive learning through physical dynamics.

cond-mat.dis-nn↗

arXiv2025

A minimal scenario for the origin of non-equilibrium order

Life uses non-equilibrium mechanisms to create ordered structures not attainable at equilibrium; the resulting order is assumed to provide functional benefits that outweigh costs of time and energy needed by these mechanisms. Here, we show that models of DNA replication and self-assembly, when expanded to include known stalling effects, can evolve error correcting mechanisms like kinetic proofreading and dynamic instability through selection for fast replication alone. We abstract these results into a general framework that predicts a counterintuitive ''order through speed'' effect if the distribution of replication times is wide enough. We test our results against recent mutational screens of proofreading polymerases. Our work suggests the intriguing possibility that non-equilibrium order can evolve even before that order is directly functional, with consequences for the evolution of mutation rates, viral capsid assembly, and the origin of life.

q-bio.PE↗

arXiv2024

Soft Modes as a Predictive Framework for Low Dimensional Biological Systems across Scales

All biological systems are subject to perturbations: due to thermal fluctuations, external environments, or mutations. Yet, while biological systems are composed of thousands of interacting components, recent high-throughput experiments show that their response to perturbations is surprisingly low-dimensional: confined to only a few stereotyped changes out of the many possible. Here, we explore a unifying dynamical systems framework - soft modes - to explain and analyze low-dimensionality in biology, from molecules to eco-systems. We argue that this one framework of soft modes makes non-trivial predictions that generalize classic ideas from developmental biology to disparate systems, namely: phenocopying, dual buffering, and global epistasis. While some of these predictions have been borne out in experiments, we discuss how soft modes allow for a surprisingly far-reaching and unifying framework in which to analyze data from protein biophysics to microbial ecology.

q-bio.QM↗

arXiv2024

Limits on the computational expressivity of non-equilibrium biophysical processes

Many biological decision-making processes can be viewed as performing a classification task over a set of inputs, using various chemical and physical processes as "biological hardware." In this context, it is important to understand the inherent limitations on the computational expressivity of classification functions instantiated in biophysical media. Here, we model biochemical networks as Markov jump processes and train them to perform classification tasks, allowing us to investigate their computational expressivity. We reveal several unanticipated limitations on the input-output functions of these systems, which we further show can be lifted using biochemical mechanisms like promiscuous binding. We analyze the flexibility and sharpness of decision boundaries as well as the classification capacity of these networks. Additionally, we identify distinctive signatures of networks trained for classification, including the emergence of correlated subsets of spanning trees and a creased "energy landscape" with multiple basins. Our findings have implications for understanding and designing physical computing systems in both biological and synthetic chemical settings.

cond-mat.dis-nn↗

arXiv2024

Learning dynamical behaviors in physical systems

Physical learning is an emerging paradigm in science and engineering whereby (meta)materials acquire desired macroscopic behaviors by exposure to examples. So far, it has been applied to static properties such as elastic moduli and self-assembled structures encoded in minima of an energy landscape. Here, we extend this paradigm to dynamic functionalities, such as motion and shape change, that are instead encoded in limit cycles or pathways of a dynamical system. We identify the two ingredients needed to learn time-dependent behaviors irrespective of experimental platforms: (i) learning rules with time delays and (ii) exposure to examples that break time-reversal symmetry during training. After providing a hands-on demonstration of these requirements using programmable LEGO toys, we turn to realistic particle-based simulations where the training rules are not programmed on a computer. Instead, we elucidate how they emerge from physico-chemical processes involving the causal propagation of fields, like in recent experiments on moving oil droplets with chemotactic signalling. Our trainable particles can self-assemble into structures that move or change shape on demand, either by retrieving the dynamic behavior previously seen during training, or by learning on the fly. This rich phenomenology is captured by a modified Hopfield spin model amenable to analytical treatment. The principles illustrated here provide a step towards von Neumann's dream of engineering synthetic living systems that adapt to the environment.

cond-mat.soft↗

arXiv2024

Training physical matter to matter

Biological systems offer a great many examples of how sophisticated, highly adapted behavior can emerge from training. Here we discuss how training might be used to impart similarly adaptive properties in physical matter. As a special form of materials processing, training differs in important ways from standard approaches of obtaining sought after material properties. In particular, rather than designing or programming the local configurations and interactions of constituents, training uses externally applied stimuli to evolve material properties. This makes it possible to obtain different functionalities from the same starting material (pluripotency). Furthermore, training evolves a material in-situ or under conditions similar to those during the intended use; thus, material performance can improve rather than degrade over time. We discuss requirements for trainability, outline recently developed training strategies for creating soft materials with multiple, targeted and adaptable functionalities, and provide examples where the concept of training has been applied to materials on length scales from the molecular to the macroscopic.

cond-mat.soft↗

arXiv2024

Dynamic coexistence driven by physiological transitions in microbial communities

Microbial ecosystems are commonly modeled by fixed interactions between species in steady exponential growth states. However, microbes often modify their environments so strongly that they are forced out of the exponential state into stressed or non-growing states. Such dynamics are typical of ecological succession in nature and serial-dilution cycles in the laboratory. Here, we introduce a phenomenological model, the Community State model, to gain insight into the dynamic coexistence of microbes due to changes in their physiological states. Our model bypasses specific interactions (e.g., nutrient starvation, stress, aggregation) that lead to different combinations of physiological states, referred to collectively as "community states", and modeled by specifying the growth preference of each species along a global ecological coordinate, taken here to be the total community biomass density. We identify three key features of such dynamical communities that contrast starkly with steady-state communities: increased tolerance of community diversity to fast growth rates of species dominating different community states, enhanced community stability through staggered dominance of different species in different community states, and increased requirement on growth dominance for the inclusion of late-growing species. These features, derived explicitly for simplified models, are proposed here to be principles aiding the understanding of complex dynamical communities. Our model shifts the focus of ecosystem dynamics from bottom-up studies based on idealized inter-species interaction to top-down studies based on accessible macroscopic observables such as growth rates and total biomass density, enabling quantitative examination of community-wide characteristics.

q-bio.PE↗

arXiv2023

Pattern recognition in the nucleation kinetics of non-equilibrium self-assembly

Inspired by biology's most sophisticated computer, the brain, neural networks constitute a profound reformulation of computational principles. Remarkably, analogous high-dimensional, highly-interconnected computational architectures also arise within information-processing molecular systems inside living cells, such as signal transduction cascades and genetic regulatory networks. Might neuromorphic collective modes be found more broadly in other physical and chemical processes, even those that ostensibly play non-information-processing roles such as protein synthesis, metabolism, or structural self-assembly? Here we examine nucleation during self-assembly of multicomponent structures, showing that high-dimensional patterns of concentrations can be discriminated and classified in a manner similar to neural network computation. Specifically, we design a set of 917 DNA tiles that can self-assemble in three alternative ways such that competitive nucleation depends sensitively on the extent of co-localization of high-concentration tiles within the three structures. The system was trained in-silico to classify a set of 18 grayscale 30 x 30 pixel images into three categories. Experimentally, fluorescence and atomic force microscopy monitoring during and after a 150-hour anneal established that all trained images were correctly classified, while a test set of image variations probed the robustness of the results. While slow compared to prior biochemical neural networks, our approach is surprisingly compact, robust, and scalable. This success suggests that ubiquitous physical phenomena, such as nucleation, may hold powerful information processing capabilities when scaled up as high-dimensional multicomponent systems.

cond-mat.dis-nn↗

arXiv2023

Curiosity-driven search for novel non-equilibrium behaviors

Exploring the spectrum of novel behaviors a physical system can produce can be a labor-intensive task. Active learning is a collection of iterative sampling techniques developed in response to this challenge. However, these techniques often require a pre-defined metric, such as distance in a space of known order parameters, in order to guide the search for new behaviors. Order parameters are rarely known for non-equilibrium systems a priori, especially when possible behaviors are also unknown, creating a chicken-and-egg problem. Here, we combine active and unsupervised learning for automated exploration of novel behaviors in non-equilibrium systems with unknown order parameters. We iteratively use active learning based on current order parameters to expand the library of known behaviors and then relearn order parameters based on this expanded library. We demonstrate the utility of this approach in Kuramoto models of coupled oscillators of increasing complexity. In addition to reproducing known phases, we also reveal previously unknown behavior and the related order parameter.

cond-mat.soft↗