PRX Life

Cephalopod skin exhibits a wide range of dynamic visual displays, including traveling waves, light/dark flashes, and flickering, all generated by muscle-regulated expansion and retraction of pigment-filled elastic sacs known as chromatophores. While central neural control is essential for high-level body-wide pattern coordination, experimental evidence from excised and denervated skin reveals that intricate spatiotemporal patterns can also emerge spontaneously in the skin. This raises questions about the respective roles of central versus peripheral control. To determine the simplest control necessary to generate such patterns, we developed ChromatoNet, a biophysical model of chromatophore arrays that includes mechanical interactions on 1- and 2-dimensional lattices with the excitability properties of radial muscles. With minimal external input, our modified Morris--Lecar model reproduces a rich repertoire of skin dynamics, including spontaneous flickering as well as traveling and spiral waves. Both spontaneous and stimulus-evoked waves emerge at an intrinsic resonant frequency, determined by model parameters. Propagation speeds remain stable across conditions and are consistent with the range observed in denervated skin preparations. The model generates diverse output states in response to local couplings and the time-dependence and shape of the input stimulation, with little top-down direction. The results provide a proof-of-principle that chromatophore/muscle dynamics with minimal central neural driving can produce rich dynamic patterns.

The emergence of treatment resistant mutants is one of the primary contributors to therapy failure in cancer. Evolution-based adaptive therapies aim to exploit the competition between sensitive and resistant cells to prolong therapeutic control. However, in the densely packed tissues of solid tumors, competition is not only determined by cell-intrinsic factors but also a result of local physical interactions between neighboring cells. The successful development of physics-informed adaptive therapies therefore hinges upon the bottom-up integration of tissue mechanics and experimentally grounded models of tumor evolution. Here, we show how population-level treatment response emerges from mechanical interactions between individual cells. Combining genetically tailored therapy-mimicry experiments in spatially structured 2D melanoma colonies with mechanistic, particle-based tissue simulations as digital twins in a real-to-sim-to-real framework, we reveal how treatment-induced changes in mechanical crowding shape a cascade of resistance expansion, spatial competitive escape, and consequent therapy failure. We then link these underlying physical mechanisms to population-level observables in an abstracted mathematical model for reinforcement-learning-based therapy optimization and a distilled treatment-decision-heuristic, which we test experimentally. Together, our results identify mechanical crowding and escape as governing principles of resistance-mediated therapy failure in solid tumours and provide a basis for mechanistically informed, evolution-based treatment strategies.

Behavioral variability reflects differences in internal and external parameters and provides a substrate for natural selection. However, uncovering its structure is challenging because behavior evolves nonlinearly across multiple timescales. Our goal is to reconstruct phenotypic variation from finite behavioral observations without prescribing the scales at which individuals should be compared. To achieve this, we introduce a multiscale dissimilarity that compares dynamics across all statistically resolvable resolutions by constructing reduced transfer operators at progressively finer scales and weighting their differences by finite-sampling uncertainty. This dissimilarity defines a geometry of dynamical phenotypes, which we reconstruct using diffusion maps. We show that the resulting phenotypic coordinates locally follow the parameter directions to which dynamics are most sensitive. We validate our approach in a stochastic model and apply it to bacterial and larval zebrafish behavior. In bacteria, we identify run speed as a major component of inter-individual variability, associated with the chemotaxis protein CheB. In zebrafish, we find that prior prey experience reshapes behavioral dynamics and phenotypic variability, with paramecia experience concentrating fish in a prey-capture-associated region of phenotypic space. Together, these results provide a framework for identifying internal and environmental variables that structure dynamical phenotypes across scales.

Collective migration is often attributed to complex cell-cell communication, yet coordinated multicellular dynamics can also emerge in organisms lacking social interactions. Continuum descriptions successfully capture such behaviors but generally rely on phenomenological transport coefficients whose microscopic origin remains elusive. Here we show that these coefficients emerge from cell-cell collisions and can be independently determined from experiments. We investigate the aerotactic dispersal of confined colonies of the solitary amoeba Acanthamoeba castellanii, in which oxygen consumption continuously reshapes the local environment to generate self-organized guidance cues. Combining quantitative videomicroscopy, oxygen mapping with phosphorescent sensing films, and oxygen-controlled microfluidics, we directly measure both the oxygen landscape and the aerotactic response governing colony expansion. We then develop a Patlak-Keller-Segel (PKS) -type drift-diffusion model whose transport coefficients are entirely determined from independent experiments and microscopic-to-macroscopic modeling approach, without adjustable parameters. The model quantitatively reproduces the observed dynamics, including the coexistence of a fast ballistic front and a trailing transient accumulation band. We show that both diffusion and aerotactic transport are modulated by the same collision-controlled mechanism through density-dependent changes in cell persistence. More generally, our results establish transport coefficients as measurable emergent properties bridging microscopic cell dynamics and macroscopic collective behavior, providing a predictive framework for oxygen-guided migration in living active matter.

Precise neural circuit formation requires growth cones to integrate multiple, sometimes competing, guidance signals into persistent yet adaptable movement. Here, we propose a theoretical framework that recasts axon guidance as a thermodynamically regulated computation. An artificial neural network, trained on in vivo genetic data and associated outgrowth patterns, maps the microstates of a localized guidance signaling network to macroscopic outgrowth behaviors. Using the trained model, we simulated axon pathfinding through extracellular gradients of molecular cues and created dynamic entropic landscapes. These results reveal that growth cones navigate high-entropy ridges that preserve plasticity. Navigation along these ridges supports persistent extension, whereas turning and branching occur near boundaries between competing entropic macrostates, consistent with transitions in the cytoskeletal machinery that controls growth-cone movement. More broadly, the framework proposes that robust biological patterning can emerge from microscopic variability when signaling networks operate near boundaries between competing behavioral states.

Biomolecular condensates are condensed assemblies of biomolecules that form through the process of liquid phase separation. Condensates in biology typically function as membraneless organelles, providing compartmentalization in the absence of a dividing lipid membrane. The molecular make-up of different condensates includes diverse multivalent proteins and nucleic acids as the primary drivers, many of them including significant fractions of intrinsically disordered regions (IDRs). To date, the sequence to phase separation relationship of IDRs has focused largely on one protein at a time, studying single-component condensate formation, or single-component partitioning into condensates. The co-phase separation and mixing of two or more IDRs is considerably more complex as both sequences can vary widely in their self- and cross-interactions, as well as their relative abundance in solution. It is unclear that a rule which predicts how a single sequence behaves will also predict what happens when two sequences are mixed. In this study, we disentangle the influence of sequence from that of composition using a set of 18 LAF-1 RGG variants that keep the same length and amino-acid composition and change only the order of the residues. This lets us vary charge patterning and, to a lesser extent, hydropathy patterning while keeping protein composition fixed. By themselves, the sequences phase separation and single-chain compaction are controlled by their degree of charge and hydropathy patterning. Within single-component condensed phases, each sequence adopts a more extended conformational ensemble, due to a more favorable, selfsolvated environment. We find that mixing two IDRs together into a condensate causes this universal scaling behavior to break, impacted by the relative interactions of the two components and overall composition of the slab. We find two different qualitative behaviors, one characterized by cooperative co-condensation when both sequences are subcritical, and the other by scaffold-client behavior when one sequence is supercritical. The scaffold-client systems generally show a high degree of demixing, while the co-condensing systems are generally quite well-mixed in the dense phase. This is surprising because even in cases where both partners have significantly different patterning parameters, they still mix. Thus, the descriptor that predicts a sequence's behavior alone can help indicate whether it will mix with or separate from a second component, but it does not fully determine the outcome on its own.

Adherent cells must continuously adapt to rapid environmental fluctuations to preserve mechanical integrity. However, conventional theoretical frameworks predominantly rely on quasi-static assumptions, thereby limiting their ability to capture the transient dynamics of cellular responses to high-rate perturbations, such as acute osmotic shocks. To address this limitation, we developed a biophysical model incorporating a Hill-type law that couples active cytoskeletal mechanics to cell-edge velocity, explaining the counterintuitive experimental observation that traction forces transiently decrease during rapid hypotonic swelling despite an increase in cell size. We further construct a kinematic-geometric phase diagram that connects classical quasi-static theories with our dynamic model. Moreover, we show that traction-force dynamics are sensitive to the loading rate and amplitude of osmotic shock. In addition, stiffness-associated simulations and supporting perturbation evidence suggest a cellular stiffness-associated recovery trend in which stiffer cells exhibit faster volume recovery driven by stronger recoil of hydrostatic pressure.

Dense, well-ordered material states of proteins form the amyloid fibers that are a hallmark of neurodegenerative disease in the brain. Beyond forming amyloid fibers, some of these proteins can also adopt other material states termed condensates which are initially liquid-like but evolve to a soft, glassy phase. Fiber growth requires a large supply of monomers and, thus, it is often speculated that fibers emerge from within a dense condensate as it ages and its microscopic dynamics slow into a glassy state. Here, we use the well-established model system Fused in Sarcoma (FUS) to directly observe, quantify and theoretically describe fiber growth and its interplay with condensates. We report the discovery that fibers grow overwhelmingly in the dilute phase surrounding the condensates while the condensates concurrently evolve to a glassy arrested solid. The resulting protein fibers and glassy condensates are both distinct solid-like phases that coexist but do not directly interconvert. Taken together, these findings reveal that there are two competitive aging pathways in FUS condensation that are linked through phase separation kinetics.

Proteins function through dynamic conformational ensembles rather than a single static structure, yet accurately modeling these ensembles remains challenging. Here, we introduce TopoFlow, an evolutionarily conditioned flow-matching framework for protein conformational ensemble generation. TopoFlow uses evolutionary representations from multiple sequence alignments (MSAs) to encode global conformational heterogeneity. It combines this with structural variables from a variational autoencoder (VAE) to capture local flexibility. These complementary representations are then adaptively fused through a conditional modulation module to construct unified conditioning features that guide protein conformational ensemble generation. On the ATLAS molecular dynamics (MD) benchmark, TopoFlow improved the Jensen Shannon metrics for pairwise distances (JS-PwD), radius of gyration (JS-Rg), and time-lagged independent components (JS-TIC) by 10.73%, 5.78% and 7.75%, respectively, relative to BioEmu. Notably, this performance generalizes to experiment-based evaluation, where TopoFlow exhibits strong agreement with experimental observables, including chemical shifts (CS) and small-angle X-ray scattering (SAXS). Ablation analyses further indicated that the evolutionary representations and structural latent variables contributed complementary information to ensemble generation. Our results suggest that TopoFlow, through its integration of evolutionary and structural latent information, offers a promising strategy for generating structurally plausible and conformationally informative protein ensembles.

Clathrin-mediated endocytosis (CME) is the primary pathway for internalization of extracellular and membrane cargo in eukaryotic cells. It is characterized by a patch of the plasma membrane invaginating and pinching off to generate a cytoplasmic vesicle. Recently, experimental evidence has begun to accumulate in support of proteins assembling into liquid-like condensates at nascent CME sites, but potential functional consequences are not understood. Given that condensate capillary forces can influence membrane bending, we constructed complementary analytical and numerical model systems to probe the mechanical influence of a protein droplet on endocytosis. Model evaluations and analysis reveal that a droplet can regulate the progression of endocytosis by stalling invagination until the droplet grows past a threshold size. After the droplet-mediated brake is released, the droplet provides a force that can drive the membrane bud toward scission. Thus, we propose a ``mechanical timer'' mechanism as a function for protein phase separation in cellular membrane trafficking.

Generative models of protein structures have accelerated scientific discovery. However, the leading structure-prediction models are trained primarily on crystal structures, limiting their ability to sample the conformations of intrinsically disordered proteins (IDPs). All-atom molecular dynamics (MD) simulations can probe the diversity of IDP conformational ensembles, but these simulations are slow and costly. Leveraging the speed of generative models and the accuracy of MD, we develop Zephyr, an SE(3) equivariant diffusion model trained on simulated IDP ensembles. The model generates the structure of a single protein sequence in two phases: a coarse-grained structure is first sampled and heavy atoms are subsequently denoised onto this structure. We evaluate our method on two fronts: ensemble accuracy and sampling efficiency. Across five test systems, the model reproduces key geometric and energetic metrics of the MD ensembles, while sampling at least 63 times faster.

We develop a general geometrical framework for growth processes in which the growth rate depends exponentially on the number of active growth-supporting components. The framework separates the fraction of active components from the size of the accessible component space, allowing qualitatively different growth laws to arise from the same underlying construction. Bacterial population growth provides an example of a fixed-size component space controlled primarily by activation and deactivation, whereas examples such as human population growth or speciation provide an expanding component space in which interactions between individuals promote further growth. This motivates a Replicative geometrical law, dX/dt= X exp({beta} X^{gamma}), whose exponent characterizes the scaling of the effective growth-supporting component space and is tested against historical human-population regimes and eukaryotic lineage-through-time data. For innovation indicators such as patents and arXiv manuscript submissions, where an existing item need not itself reproduce, the same construction motivates a Direct Geometrical production law without the prefactor X. The framework therefore provides a common description of fixed and expanding component spaces while distinguishing Replicative from Direct Geometrical growth.

Deep learning co-folding and docking models accurately predict ligand binding poses at orthosteric sites yet systematically underperform at allosteric pockets. Here, we demonstrate that this accuracy gap reflects a fundamental biophysical property of local energy landscape topology specifically, the magnitude of frustration quenching upon ligand binding rather than an intrinsic algorithmic defect. By evaluating leading co-folding and docking paradigms across curated benchmarks and large-scale uncurated datasets, we show that AI prediction success strongly tracks frustration quenching which the degree to which ligand binding relieves local energetic strain to form a steep energy funnel. Topologically, frustration-weighted residue interaction networks reveal that orthosteric pockets act as high-betweenness communication hubs embedded within contiguous, minimally frustrated pathways. In contrast, allosteric sites reside in low-betweenness peripheral regions with fragmented frustration profiles, presenting flatter, entropy-dominated landscapes that challenge structure prediction algorithms. Comparing generative diffusion with single-shot regression demonstrates that multi-pose sampling capacity is required to cross energy barriers and discover the frustration funnel, whereas joint receptor-ligand coordinate co-adaptation governs local optimization within the funnel. Collectively, this study reframes the allosteric "blind spot" as a quantitative property of energy landscape architecture rather than a defect of any specific predictive framework. We demonstrate that frustration quenching magnitude and structural network centrality serve as physical descriptors that define the intrinsic limits of AI performance, reframing AI accuracy from a benchmark metric into a physics-grounded quantitative reporter of macromolecular binding landscape architecture.

Cells use ATP-driven chaperones to extract clients from biomolecular condensates selectively, yet no closed-form criterion links molecular parameters to condensate quality control. We derive a minimal non-equilibrium sorting reactor model that reduces the reaction-diffusion dynamics to one algebraic equation. In the idealized low-load limit, a concentration-dependent partition coefficient produces a saddle-node bifurcation between functional clearance and pathological storage. The critical Damk\"ohler number is \(\mathrm{Da}_{0c}=e^{-2}\) and is independent of the feedback strength. Corrections shift the value but preserve the topology. The model predicts a Kramers-type functional lifetime and recovery costs that scale as \(R^3\). Critical slowing down near the bifurcation provides early-warning signatures for FCS and FRAP.

Phase separating biopolymers are often described by the sticker-spacer model, which accounts for non-attractive linkers that connect discrete attractive moieties. "Associative polymer" theories describe these systems as two coexisting fluids of stickers and spacers, respectively. In such theories the stickers and spacers are decoupled in the sense that sticker binding does not affect spacer conformations or vice versa. Informed by coarse-grained simulations, we develop a phenomenological mean-field theory to determine the microscopic coupling between the sticker and spacer degrees of freedom. Accounting for linkers connecting adjacent spacers introduces free energy contributions due to conformational entropy of the spacer and the binding combinatorics of the stickers. These contributions have opposite dependencies on spacer length because longer spacers allow a greater volume to search for binding partners but incur a greater entropy cost upon bond formation. The length dependent terms from these contributions cancel leaving a length-independent elastic term arising from the fact that a small amount of spacer stretching lowers the free energy by increasing the number of possible binding partners. The cancellation of spacer length from the theory justifies the decoupling approximation and supports mean-field treatments of heterogeneous spacers.

Large shear deformation in cartilage can lead to tissue failure and injury, which are often precursors to diseases such as post-traumatic osteoarthritis. Under shear, cartilage strain-stiffens prior to failure, which makes stiffening an important indicator of changes to the loading of the extracellular matrix that gives rise to this tissue's properties. While a variety of stiffening mechanisms ranging from uncrimping to fiber-alignment have been explored, whether such mechanisms apply to the depth dependent shear strain stiffening in articular cartilage remains poorly understood. Here, we present experiments and theory that support the idea that the strain-stiffening behavior is caused by a mechanical phase transition from a random isotropic network whose mechanics arise primarily from bending-dominated modes to one where the mechanics arise from stretching-dominated modes. Indeed, consistent with network model predictions, we find that stress-strain data for all depths can be collapsed onto a single universal curve that diverges at a depth-dependent critical strain. These results show that increasing the local concentration of collagen, which constitutes the primary load-bearing network in cartilage, decreases the critical strain at which the strain-stiffening transition occurs, while increasing the low-strain elastic modulus. Furthermore, we demonstrate that bulk compression leads to a higher critical strain and decreased low-strain elastic modulus. Network simulations capture this rich behavior, and indicate that under compression the network collapses, which allows for larger strains before the transition from bending-dominated to stretching-dominated mechanics occurs. Collectively, these findings shed light on the roles that tissue composition and axial strain play in the mechanisms that organize how articular cartilage stiffens during the lead-up to injury.

Hox proteins are evolutionary conserved transcription factors guiding morphogenesis in all bilaterian animals. Conservation mainly concerns the homeodomain (HD), a helix-turn-helix DNA binding motif and flanking sequences close by. Besides, sequences are poorly conserved and disordered. Here we investigate how poorly conserved Intrinsically Disordered Regions (IDRs) influence Hox protein activity, focusing on a closely related Drosophila Hox pair and using a wide range of in vivo function. Studying Drosophila Ultrabithorax (Ubx) and AbdominalA (AbdA) proteins, we find that conserved protein motifs close to the HD are differently used in Ubx and AbdA. Coarse-grained molecular dynamics simulations identify a large set of intra-molecular interactions, notably involving poorly conserved disordered regions upstream and downstream the HD. Swapping these regions in Ubx and AbdA alters protein functions and switches motif dependency in some biological situations. This suggests that IDRs, predicted to mediate large and dynamic sets of intramolecular interactions, influence protein motif usage in a context dependent manner.

Understanding how conformational dynamics regulate protein complex assembly remains a central challenge in molecular systems biology. Here, we develop a multiscale modeling framework that integrates all-atom molecular dynamics (MD) simulations with rule-based kinetic modeling to investigate the formation of the PP1-GADD34-eIF2 complex, a critical regulator of the integrated stress response. Using umbrella sampling, we compute the potential of mean force (PMF) for key interactions, revealing strong thermodynamic driving forces for eIF2 binding ({Delta}G = -89.5 kJ/mol, KD {approx} 0.07 fM) and moderate affinity for PP1-GADD34 association ({Delta}G = -40.7 kJ/mol, KD {approx} 0.138 M). These energetics inform a PySB-based model that incorporates GADD34's folding state as a continuous variable ({varphi}), linking conformational transitions to holoenzyme assembly and activity. Simulations show that catalytic efficiency is maximized when GADD34 folding free energy is 20 kJ/mol and {varphi} = 0.9, reflecting a highly ordered state stabilized by actin and PP1 binding. Comparative analysis of energy-constrained and optimized parameter regimes reveals that folding-dependent assembly enhances both eIF2 dephosphorylation and information transmission, with channel capacities increasing from <0.002 bits to >5.5 bits. These results demonstrate how conformational regulation encodes system-level function and underscore the utility of multiscale models in bridging molecular energetics with dynamic biochemical control.

Protein language models have become a dominant foundation for sequence-based protein-protein interaction (PPI) prediction, but their generalization remains limited under stringent evaluation, and benchmark accuracy alone cannot reveal whether a PPI predictor learns partner-specific biological signals or exploits contextual shortcuts. Here we introduce AuditPPI, an interpretable framework that transforms sparse-autoencoder (SAE) features from a frozen protein language model into order-invariant pair fingerprints, recasting PPI prediction as an auditable tabular-learning problem. AuditPPI achieves competitive predictive performance across diverse PPI benchmarks while enabling multiscale auditing of the information supporting its predictions. Protein-level audits show that partner-independent participation signals remain substantial in conventional benchmarks, including protein-disjoint benchmark, but are largely insufficient for pair-level discrimination in topology- and degree-controlled benchmark. Pair-level analyses further show that removing protein identity overlap does not eliminate contextual structure: protein-disjoint benchmarking retains strong subcellular co-localization signals and benchmark-dependent feature matching captured by SAE co-activation and absolute-difference. Structural analyses, however, provide little evidence that these predictive signals are specifically grounded in PPI interfaces: feature enrichment at interfaces for protein-disjoint benchmark was no longer detectable after controlling for surface exposure, and contact-specific enrichment was observed for only a small fraction of testable SAE feature pairs. AuditPPI therefore separates predictive success from mechanistic fidelity and provides a practical framework for evaluating whether sequence-based PPI predictors capture partner-specific biological signals instead of contextual shortcuts.

We investigate the dynamics of the spread of the infection due to a phage- during the range expansion of an E coli colony. We find a rich panoply of spatio-temporal dynamics at large colony-level length scales, arising due to the interplay of mechanical forces, growth and viral infection dynamics at smaller cellular length scales. Strikingly, the phages "surf" the front of the growing bacterial colony by hitchhiking on E coli that are advected due to the growth of bacteria in the colony, resulting in an anisotropic spread of the phages. We identify microscopic processes - of the phage release during lysis (occurring on short millisecond timescales at length scales close to that of a single bacterium) and local nematic alignment of the rod-like E coli bacteria (occurring on timescales comparable to the bacterial growth rate and the length scales of a few bacterial cells) - that enhance the advective effects driving the hitchhiking behavior. Combining our experiments with a mathematical model, we explain the multiple spatio-temporal dynamical regimes - from the coexistence of uninfected bacteria, resistant (lysogenic) cells and phages, to fixed points where the entire population turns resistant or remains uninfected. Our work broaches a new frontier in the investigation of eco-evolutionary processes arising from the mechanical and ecological forces due to the interactions between viruses and their hosts.