Thermodynamic irreversibility is a fundamental characteristic of living matter, crucial for maintaining the complex spatiotemporal structures and functions that define biological systems. However, accurately quantifying biologically relevant irreversibility remains challenging due to noise and nonlinear interactions among many degrees of freedom. Here, we employ deep learning techniques to identify tractable, low-dimensional representations of phase-field patterns in a canonical protein signaling process—the Rho-GTPase system—as well as in complex Ginzburg-Landau dynamics. We demonstrate that factorizing variational autoencoder neural networks can robustly extract informative pattern features despite noise. These neural-network-derived representations reveal signatures of mesoscopic broken detailed balance and time-reversal asymmetry in both Rho-GTPase and complex Ginzburg-Landau wave dynamics. By applying the compression-based Ziv-Merhav estimator to these representations, we successfully recover irreversibility trends across complex Ginzburg-Landau patterns, which vary widely in spatiotemporal frequency and noise level. Our irreversibility estimates also recapitulate cell-activity trends in Rho-GTPase patterns subjected to metabolic inhibition. Furthermore, our framework distinguishes between stable and chaotic dynamical phases in these nonlinear systems. By leveraging advances in deep learning, our approach provides robust, model-free measurements of nonequilibrium and nonlinear behavior in complex living processes.

Recent advances in deep generative methods have allowed antibody sequence and structure co-design. This study addresses the challenge of tailoring the highly variable complementarity-determining regions (CDRs) in antibodies to fulfill developability requirements. We introduce a guidance approach that integrates property information into the antibody design process using diffusion probabilistic models. This approach allows us to simultaneously design CDRs conditioned on antigen structures while considering critical properties like solubility and folding stability. Our property-guided diffusion model offers versatility by accommodating diverse property constraints, presenting a promising avenue for computational antibody design in therapeutic applications.

One of the planet's more understudied ecosystems is the deep biosphere, where organisms can experience high hydrostatic pressures (30–110 MPa); yet, by current estimates, these subsurface and deep ocean zones host the majority of the Earth's microbial and animal life. The extent to which terrestrially relevant pressures up to 100 MPa deform most globular proteins—and which kinds—has not been established. Here, we report the invention of an experimental apparatus that enables structural proteomic methods to be carried out at high pressures for the first time. The method, called high-pressure limited proteolysis (Hi-P LiP), involves performing pulse proteolysis on whole cell extracts brought to high pressure. The resulting sites of proteolytic susceptibility induced by pressure are subsequently read out by sequencing the peptide fragments with tandem liquid chromatography–mass spectrometry. The method sensitively detects pressure-induced structural changes with residue resolution and on whole proteomes, providing a deep and broad view of the effect of pressure on protein structure. When applied to a piezosensitive thermophilic bacterium, Thermus thermophilus, we find that approximately 40% of its soluble proteome is structurally perturbed at 100 MPa. Proteins with lower charge density are more resistant to pressure-induced deformation, as expected; however, contrary to expectations, proteins with lower packing density (i.e., more voids) are also more resistant to deformation. Furthermore, high pressure has previously been shown to preferentially alter conformations around active sites. Here, we show this is also observed in Hi-P LiP, suggesting that the method could provide a generic and unbiased modality to detect binding sites on a proteome scale. Hence, data sets of this kind could prove useful for training emerging artificial intelligence models to predict cryptic binding sites with greater accuracy.

Multiphoton laser scanning microscopy is a powerful tool for deep imaging of thick biological samples. Image scanning microscopy (ISM) has demonstrated significant improvements in the signal-to-noise ratio in confocal laser scanning microscopy, while at the same time improving upon the effectively attainable resolution. Two-photon excitation (2PE), combined with ISM, has been shown to allow for deep tissue imaging with enhanced resolution compared to 2PE microscopy. Three-photon excitation (3PE) has enabled record imaging depth and contrast for multiphoton imaging, due to the superior suppression of out-of-focus signal generation. In this paper, we demonstrate super-resolution 3PE ISM. This is achieved using a single-photon avalanche detector array, and 1040-nm pulses for 3PE of blue fluorescence. This method enables subdiffraction limited resolution imaging of biological samples stained with blue fluorescent markers, such as mouse myocardial and spinal cord tissues stained with $4^{\prime },6$-diamidino-2-phenylindole. Deconvolution improves the resolving power further and allows for imaging with better than $\lambda /8$ resolution with respect to the 3PE wavelength $\lambda$. With the ISM pixel reassignment procedure, we demonstrate a resolution enhancement of $\sim 1.6$ laterally, compared to the resolution attained using a photomultiplier tube in a non-descanned detection arrangement, and a factor of $\sim 1.8$ enhancement in axial resolution. The experimentally measured three-dimensional point spread function volume is shrunk $\sim 4.4$-fold, which is close to the theoretically expected enhancement.

Biological systems often choose actions without an explicit reward signal, a phenomenon known as intrinsic motivation. The computational principles underlying this behavior remain poorly understood. In this study, we investigate an information-theoretic approach to intrinsic motivation, based on maximizing an agent's empowerment (the mutual information between its past actions and future states). We show that this approach generalizes previous attempts to formalize intrinsic motivation, and we provide a computationally efficient algorithm for computing the necessary quantities. We test our approach on several benchmark control problems, and we explain its success in guiding intrinsically motivated behaviors by relating our information-theoretic control function to fundamental properties of the dynamical system representing the combined agent-environment system. This opens the door for designing practical artificial, intrinsically motivated controllers and for linking animal behaviors to their dynamical properties.

Cells regulate fates and complex body plans using spatiotemporal signaling cascades that alter gene expression. Short DNA sequences, known as enhancers (50–1500 base pairs), help coordinate these cascades by attracting regulatory proteins that enhance the transcription by binding to distal gene promoters. In humans, there are hundreds of thousands of enhancers dispersed across the genome, which poses a challenging coordination task to prevent unintended gene activation. To mitigate this problem, the genome contains insulator elements that block enhancer-promoter interactions. However, there is an open problem with how the insulation works, especially as enhancer-insulator pairs may be separated by millions of base pairs. Based on recent empirical data from Hi-C experiments, this paper proposes a new mechanism that challenges the common paradigm that rests on specific insulator-insulator interactions. Instead, this paper introduces a stochastic looping model where insulators bind weakly to chromatin rather than other insulators. After calibrating the model to experimental data, we use simulations to study the broad distribution of hitting times between an enhancer and a promoter when insulators are present. We find parameter regimes with large differences between average and most probable hitting times. This makes it difficult to assign a typical timescale and hints at highly defocused regulation times. We also map our computational model onto a resetting problem that allows us to derive several analytical results. Besides offering new insights into enhancer-insulator interactions, our paper advances the understanding of gene regulatory networks and causal connections between genome folding and gene activation.

Homing–the incredible ability of animals to navigate back to their homes from unfamiliar places is surprisingly widespread and crucial for their survival. However, the physical understanding of this phenomenon is not yet developed. Here we use a light-controlled robot mimicking foraging and homing behavior to investigate this phenomenon. The robot as a forager is a self-propelled active particle programed to undergo an in-plane active Brownian (AB) motion whose velocity vector $\mathbf{v}$ undergoes rotational diffusion of magnitude $D_r$. During the homing phase, the robot undergoes guided motion toward a positive light gradient aided by repeated reorientations, directing it back to its home. Our key finding is an interesting optimal behavior where the mean homing time becomes independent of $D_r$ beyond a critical value posited as a signature of enhanced efficiency. We develop a first-passage-based theoretical model of homing motion, which elucidates this finding as well as accurately captures quantitative features of the homing trajectories in the form of temporal autocorrelation function of the robot's orientation. Inspired by the paradigm of stochastic resetting processes, we also perform an alternative homing motion in a computer, which integrates an AB motion with course correction resets, corroborating our experimental findings. Finally, we test our model on the publicly available data on homing pigeons and capture similar key characteristics of the homing trajectories. Together, these results offer valuable insights into the physics of homing dynamics, providing a statistical basis for its robustness across the animal kingdom.

Collective cell dynamics play a crucial role in many developmental and physiological contexts. While two-dimensional (2D) cell migration has been widely studied, how three-dimensional (3D) geometry and topology interplay with collective cell behavior to determine dynamics and functions remains an open question. In this work, we elucidate the biophysical mechanism underlying rotation in spherical tissues, a phenomenon widely reported both in vivo and in vitro. Using murine pancreas-derived organoids as a model system, we find that epithelial spheres exhibit persistent rotation, rotational axis drift, and rotation arrest. Using a 3D vertex model, we demonstrate how the combined action of traction force and polarity alignment can account for these distinct rotational dynamics near a solid to flow transition. Furthermore, our analysis shows that the spherical tissue rotates as an active solid occasionally switching to a flowing state and exhibits spontaneous chiral symmetry breaking. Using a continuum model, we demonstrate how the topological defects in the polarity field underlie this symmetry breaking process, which is revealed by asymmetries in the cell elongation pattern. For cell elongation to reveal the chiral asymmetry, shear flow is required in addition to the solid body rotation. Altogether, our work reveals a robust chiral symmetry breaking mechanism with potential implications for left-right symmetry breaking processes in morphogenetic events.

Network-forming organisms, like fungi and slime molds, dynamically reorganize their networks during foraging. The resulting rerouting of resource flows within the organism's network can significantly impact local ecosystems. In current analysis limitations stem from a focus on single-time-point morphology, hindering understanding of continuous dynamics and underlying constraints. Here we study ongoing network reorganization in the foraging slime mold Physarum polycephalum, identifying three distinct states with varying morphology and migration velocity. We estimate the energetic cost of each state and find a trade-off between building and transport costs within the morphological variability, facilitating different search strategies. Adaptation of state population to the environment suggests that diverse network morphologies support varied foraging strategies, though constrained by associated costs. Our findings provide insights for evaluating the impact of resource flow rerouting in changing ecosystems.

Bacteria thrive in anisotropic media such as biofilms, biopolymer solutions, and soil pores. In strongly mechanically anisotropic media, physical interactions force bacteria to swim along a preferred direction rather than to execute the three-dimensional random walk due to their run-and-tumble behavior. Despite their ubiquity in nature and importance for human health, there is little understanding of bacterial mechanisms to navigate these media while constrained to one-dimensional motion. Using a biocompatible liquid crystal, we discovered two mechanisms used by Bacillus subtilis bacteria to switch directions in anisotropic media. First, the flagella assemble in bundles that work against each other from opposite ends of the cell body, and the dominating side in this flagellar “Tug-of-Oars” propels the bacterium along the nematic direction. Bacteria frequently revert their swimming direction ${180}^{\circ }$ by a mechanism of flagellar buckling and reorganization on the opposite side of the cell. The Frank elastic energies of the liquid crystal dictate the minimum compression for the Euler buckling of a flagellum. Beyond a critical elasticity of the medium, flagellar motors cannot generate the necessary torque for flagellar buckling, and bacteria are stuck in their configuration. However, we found that bacteria can still switch swimming directions using a second mechanism where individual bundles alternate their rotation. Our results shed light on bacterial strategies to navigate anisotropic media and give rise to questions about sensing environmental cues and adapting at the level of flagellar bundles. The two adaptation mechanisms found here support the use of biocompatible liquid crystals as a synthetic model for bacterial natural environments.

Understanding the mechanisms governing the structure and dynamics of flexible polymers like chromosomes, especially the signatures of motor-driven active processes, is of great interest in genome biology. We study chromosomes as a coarse-grained polymer model where microscopic motor activity is captured via an additive temporally persistent noise. The active steady state is characterized by two parameters: active force, controlling the persistent-noise amplitude, and correlation time, the decay time of active noise. We find that activity drives correlated motion over long distances and a regime of dynamic compaction into a globally collapsed entangled globule. Diminished topological constraints destabilize the entangled globule, and the active segments trapped in the globule move toward the periphery, resulting in an enriched active monomer density near the periphery. We also show that heterogeneous activity leads to the segregation of the highly dynamic species from the less dynamic one, suggesting a role of activity in chromosome compartmental segregation. Adding activity to experimental-data-derived structures, we find active loci may mechanically perturb and switch compartments established via epigenetics-driven passive self-association. The key distinguishing signatures of activity are enhanced apparent diffusivity, exploration of all the dynamic regimes (subdiffusion, effective diffusion, and superdiffusion) at various lag times, and a broadened distribution of observables like the dynamic exponents.

Cellular actin dynamics results from the collective action of hundreds of regulatory proteins, majority of which target actin filaments at their barbed ends. Three key actin binding proteins—profilin, cofilin, and twinfilin—individually depolymerize filament barbed ends. Notwithstanding recent leaps in our understanding of their individual action, how they collectively regulate filament dynamics remains an open question. In the absence of direct and simultaneous visualization of these proteins at barbed ends, gaining mechanistic insights has been challenging. We have here investigated multicomponent dynamics of profilin, cofilin, and twinfilin using a hybrid approach that combines high-throughput single filament experiments with theory. We discovered that while twinfilin competes with profilin, it promotes binding of cofilin to filament sides. Interestingly, contrary to previous expectations, we found that profilin and cofilin can simultaneously bind the same filament barbed end, resulting in its accelerated depolymerization. Our study reveals that pairwise interactions can effectively capture depolymerization dynamics in simultaneous presence of all three proteins. We thus believe that our approach of employing a theory-experiment dialog can potentially help decipher multicomponent regulation of actin dynamics.

The mesoscale organization of molecules into membraneless biomolecular condensates is emerging as a key mechanism of rapid spatiotemporal control in cells. Principles of biomolecular condensation have been revealed through in vitro reconstitution. However, intracellular environments are much more complex than test-tube environments: they are viscoelastic, highly crowded at the mesoscale, and are far from thermodynamic equilibrium due to the constant action of energy-consuming processes. We developed synDrops, a synthetic phase separation system, to study how the cellular environment affects condensate formation. Three key features enable physical analysis: synDrops are inducible, bioorthogonal, and have well-defined geometry. This design allows kinetic analysis of synDrop assembly and facilitates computational simulation of the process. We compared experiments and simulations to determine that macromolecular crowding promotes condensate nucleation but inhibits droplet growth through coalescence. ATP-dependent cellular activities help overcome the frustration of growth. In particular, stirring of the cytoplasm by actomyosin dynamics is the dominant mechanism that potentiates droplet growth in the mammalian cytoplasm by reducing confinement and elasticity. Our results demonstrate that mesoscale molecular assembly is favored by the combined effects of crowding and active matter in the cytoplasm. These results move toward a better predictive understanding of condensate formation in vivo.

Phase-separated biomolecular condensates containing proteins and RNAs can assemble into higher-order structures by forming thermodynamically stable interfaces between immiscible phases. Using a minimal model of a protein/RNA interaction network, we demonstrate how a “shared” protein species that partitions into both phases of a multiphase condensate can function as a tunable surfactant that modulates the interfacial properties. We use Monte Carlo simulations and free-energy calculations to identify conditions under which a low concentration of this shared species is sufficient to trigger a wetting transition. We also describe a numerical approach based on classical density functional theory to predict concentration profiles and surface tensions directly from the model protein/RNA interaction network. Finally, we show that the wetting phase diagrams that emerge from our calculations can be understood in terms of a simple model of selective adsorption to a fluctuating interface. Our work shows how a low-concentration protein species might function as a biological switch for regulating multiphase condensate morphologies.

Biological cells and small organisms navigate in concentration fields of signaling molecules using two fundamental gradient-sensing strategies: spatial comparison of concentrations measured at different positions on their surface and temporal comparison of concentrations measured at different locations visited along their motion path. It is believed that size and speed dictate which gradient-sensing strategy cells choose, yet this has never been formally proven. Using information theory, we investigate the optimal gradient-sensing mechanism for an ideal chemotactic agent that combines spatial and temporal comparisons. We account for the physical limits of chemosensation: molecule counting noise at physiological concentrations and motility noise inevitable at the microscale. Our simulation data collapse onto an empirical power law that predicts an optimal weighting of information as a function of motility and sensing noise, demonstrating how spatial comparison becomes more beneficial for agents that are large, slow, and less persistent. This refines and quantifies the previous heuristic notion. Our idealized model assuming unlimited information processing capabilities serves as a benchmark for the chemotaxis of biological cells.

Biomolecular condensates are dynamic intracellular entities defined by their sequence- and composition-encoded material properties. During aging, these properties can change dramatically, potentially leading to pathological solidlike states, the mechanisms of which remain poorly understood. Recent experiments reveal that the aging of condensates involves a complex interplay of solvent depletion, strengthening of sticker links, and the formation of rigid structural segments such as beta fibrils. In this study, we use various coarse-grained models to investigate how solvent expulsion, biopolymer chain rigidity, and the lifetimes of sticker contacts influence the viscoelastic properties and aging dynamics of condensates. We find that the rigidity of the biopolymer backbone is essential for replicating the predominant elastic behavior observed in experiments. In contrast, models using fully flexible chains—an assumption common in simulations of intrinsically disordered proteins—fail to exhibit a dominant elastic regime. We also demonstrate that altering the solvent content within condensates affects the crossover between storage and loss moduli. This suggests that desolvation plays a significant role in condensate aging by promoting the transition from a viscous to an elastic state. Furthermore, the lifetime of sticker pairs profoundly influences the mature state of the condensates; short-lived stickers lead to a Maxwell fluid behavior, while longer-lived, irreversibly cross-linked stickers result in solidlike properties, consistent with the Kelvin-Voigt model. Finally, by incorporating the chain rigidification, desolvation, and sticker pair formation into a nonequilibrium dynamic aging simulation, we show the molecular mechanism of forming solid shells around the condensate surfaces observed in a recent experimental report.

The evolution of resistance remains one of the primary challenges for modern medicine, from infectious diseases to cancers. Many of these resistance-conferring mutations often carry a substantial fitness cost in the absence of treatment. As a result, we would expect these mutants to undergo purifying selection and be rapidly driven to extinction. Nevertheless, preexisting resistance is frequently observed from drug-resistant malaria to targeted cancer therapies in non-small-cell lung cancer (NSCLC) and melanoma. Solutions to this apparent paradox have taken several forms, from spatial rescue to simple mutation supply arguments. Recently, in an evolved resistant NSCLC cell line, we found that frequency-dependent ecological interactions between ancestor and resistant mutant ameliorate the cost of resistance in the absence of treatment. Here, we hypothesize that frequency-dependent ecological interactions in general play a major role in the prevalence of preexisting resistance. We combine numerical simulations with robust analytical approximations to provide a rigorous mathematical framework for studying the effects of frequency-dependent ecological interactions on the evolutionary dynamics of preexisting resistance. First, we find that ecological interactions significantly expand the parameter regime under which we expect to observe preexisting resistance. Next, even when positive ecological interactions between mutants and ancestors are rare, these resistant clones provide the primary mode of evolved resistance because even weak positive interaction leads to significantly longer extinction times. We then find that even in the case where mutation supply alone is sufficient to predict preexisting resistance, frequency-dependent ecological forces still contribute a strong evolutionary pressure that selects for increasingly positive ecological effects (negative frequency-dependent selection). Finally, we genetically engineer several of the most common clinically observed resistance mechanisms to targeted therapies in NSCLC, a treatment notorious for preexisting resistance. We find that each engineered mutant displays a positive ecological interaction with their ancestor. As a whole, these results suggest that frequency-dependent ecological effects can play a crucial role in shaping the evolutionary dynamics of preexisting resistance.

Indirect reciprocity is a mechanism that explains large-scale cooperation in human societies. In indirect reciprocity, an individual chooses whether to cooperate with another based on reputation information, and others evaluate the action as good or bad. Under what evaluation rule (called “social norm”) cooperation evolves has long been of central interest in the literature. It has been reported that if individuals can share their evaluations (i.e., public reputation), social norms called “leading eight” can be evolutionarily stable. On the other hand, when they cannot share their evaluations (i.e., private assessment), the evolutionary stability of cooperation is still in question. To tackle this question, we create a novel method to analyze the reputation structure in the population under private assessment. Specifically, we characterize each individual by two variables, “goodness” (what proportion of the population considers the individual as good) and “self-reputation” (whether an individual thinks of him or herself as good or bad), and analyze the stochastic process of how these two variables change over time. We discuss the evolutionary stability of each of the leading-eight social norms by studying the robustness against invasions of unconditional cooperators and defectors. We identify key pivots in those social norms for establishing a high level of cooperation or stable cooperation against mutants. Our finding gives an insight into how human cooperation is established in a real-world society.

The presumed proximity to a critical point is believed to endow the brain with scale-invariant statistics, which are thought to confer various functional advantages in terms of its information processing, storage, and transmission capabilities. To assess the relationship between scaling and cortical states, we apply a phenomenological renormalization group analysis to 3-h spiking data recordings from the urethane-anesthetized rat's visual cortex. Under this type of anesthesia, cortical states dynamically shift across a spectrum of synchronization levels, defined by population spiking rate variability. By developing a scaling criterion based on the kurtosis of the momentum-space activity distribution, our study combines the coarse-graining method with state-dependent analysis. We find that scaling signatures only appear as spiking variability surpasses a specified threshold. Notably, within this regime, scaling exponents show relative stability. Conversely, subthreshold activity is primarily asynchronous and fails to meet the scaling criterion. Our results suggest that a wide range of cortical states corresponds to small deviations around a critical point, with the system fluctuating in and out of criticality, spending roughly three-quarters of the experiment duration within a scaling regime.