Papers

Whole slide image analysis is commonly formulated as multiple instance learning (MIL), where instance features are contextually updated and aggregated into a slide representation, a process we term slide encoding dynamics. Recently, selective state-space models (SSM) have emerged as promising MIL architectures due to their long-sequence modeling capability and linear complexity. However, existing SSM-based MIL methods rely solely on visual features during MIL. Meanwhile, in large-scale WSIs, where sparse diagnostically decisive regions are surrounded by abundant irrelevant information, such purely vision-driven selective dynamics can misallocate state updates and readouts, causing the evolving SSM state to accumulate task-irrelevant evidence and dilute critical diagnostic cues over long scan trajectories. In this work, we propose the Knowledge-Aware Hidden-State Modulation architecture (KHiM-Mamba), which innovatively regulates Mamba's core selective state-space mechanism with explicit knowledge priors, steering slide encoding dynamics toward diagnostically meaningful evidence accumulation. Specifically, we redesign the original SSM layer to perform knowledge modulation operations during the evolution of hidden states, thereby guiding what visual evidence is accumulated and retrieved from the hidden state at each encoding step. Furthermore, we additionally introduce a local-adaptive vocabulary retrieval module that uses large language models to assign each patch fine-grained, tissue-specific semantic descriptions, enabling precise modulation across diverse tasks. Experiments on 11 public benchmarks across 4 tasks show that KHiM-Mamba consistently achieves state-of-the-art performance.

Cellular activity drives epithelial fluidization — a widespread phenomenon observed during tissue development, remodeling, and repair both in vivo and in vitro. Yet the physical origins and spatial organization of active forces vary widely across biological systems and are often represented by a single generic mechanism in theoretical models. Here, using an active vertex model, we systematically compare four modes of epithelial activity spanning subcellular to tissue scales: apolar motility, polar motility, fluctuating contractility, and mechanochemical regulation. Although all four mechanisms drive the same global transition from a solid-like rectangular tissue to a fluid-like circular morphology, they reach this state through distinct pathways — differing in the rates and topology of junctional rearrangements, cell elimination, and collective motion and leave distinguishable signatures in tissue architecture, cell dynamics, and mechanical relaxation. Among these observables, spatial velocity correlations directly capture the spatial organization of activity: their correlation length and functional form together resolve all four mechanisms. The robustness of these signatures across activity strengths suggests that spatial velocity correlations offer an experimentally accessible means of identifying the physical origin of epithelial activity from live-cell imaging alone.

Extinction is inevitable; every species eventually dies out, impacting the ecosystem it is part of. Over the past few decades, extensive research stemming from the stability-diversity debate has addressed how species diversity contributes to the stability of large ecosystems. However, conventional stability criteria often rely on deterministic frameworks, overlooking the intrinsic population fluctuations that allow any species to go extinct by chance. In this paper, we incorporate demographic stochasticity into large complex ecosystems with a rule-based model. We demonstrate that such demographic fluctuations rapidly prune the low-abundance species from the ecological community, thereby securing higher systemic stability. By developing a bottom-up theory to characterise the statistics of extinction dynamics, we discover that the fraction of surviving species exhibits an anomalous heavy-tailed decay over time, revealing the emergence of remnant communities with robust metastability. Our results highlight that when demographic fluctuations are accounted for, ecosystems self-stabilise by reducing their diversity even when they are predicted to be chaotic in the deterministic limit.

Bioinformatics workflows rely heavily on visual representations. Quality-control plots, cell embeddings, heatmaps, genome-browser tracks, and interactive dashboards are not merely illustrations, but instruments for making analytical decisions. For blind and low-vision researchers who use screen readers, braille displays, or audio-based interfaces, these create a barrier: the evidence used to justify an analysis is often encoded in visual form, while the underlying decision remains undocumented. We argue that non-visual accessibility and computational reproducibility are closely aligned, as they both require analyses to be transparent and to record why decisions were made. We present ten simple rules for non-visual bioinformatics, covering plots as decision records, cautious use of AI-generated figure descriptions, accessible computing environments, text-first literate programming, structured data and metadata, compact object summaries, accessible publication formats, collaboration practices, shared community infrastructure, and accessibility as part of FAIR research. The intended audience is computational biologists and developers. Using single-cell RNA-seq as a running example, we show that the accessible equivalent of a plot is a structured decision record. That is, a plot companion that goes beyond storing the underlying data by also stating the purpose of the analysis and the resulting quantitative evidence and uncertainty. We argue that treating accessibility in this way makes bioinformatics more inclusive and also more transparent and auditable.

What decides whether people do things together or separately? Many activities cannot be carried out alone, and an individual must rank them against the private tasks competing for the same time. We address this within the priority-queue description of human activity by letting each agent choose the priority of a shared task rather than drawing it from a fixed distribution: the value of the joint activity, discounted by the estimated risk that the partner will not take part. Participation becomes strategic, and the model acquires a phase transition. A coupled phase, in which joint activity is sustained, is separated from an absorbing solitary phase by a saddle-node bifurcation that we obtain in closed form. The transition is discontinuous and the solitary phase is absorbing, so collapse is irreversible unless an agent persists unilaterally for of order one memory time, a cost we also compute. The heavy-tailed interevent statistics that motivate queueing models of human dynamics survive only in a narrow window at the transition, and there the exponent is fixed by the fraction of time spent coupled rather than by queue length: the universality classes of the non-strategic model do not survive the introduction of choice. On a network, attention divides as $1/(k+a)$ and fixes a critical degree beyond which no coupled state exists, so the solitary phase percolates according to the Molloy–Reed criterion with the second moment truncated at that degree — formally an attack on hubs, with no attacker. For group activities the critical degree falls as the $(m-1)$th root, implying a maximum group size. The transition organises two quantities already measured in communication records: a finite capacity for keeping ties active, and the decay of ties whose rhythm is interrupted.

Multimodal Large Language Models (MLLMs) can generate pathological descriptions from histological images, but gigapixel Whole Slide Images (WSIs) exceed their visual context limits. The standard tiling workaround makes WSIs tractable yet severs the tissue neighborhoods that define tumor-stroma interfaces and morphology. We introduce Spatial Language Message Passing (SLMP), a framework that performs spatial reasoning entirely in language space, human-readable by construction. SLMP represents a WSI region as a spatial text graph: tiles are nodes initialized with MLLM descriptions, and edges encode spatial adjacency. For each tile, an LLM refines its description by integrating language messages from adjacent tiles under a shared aggregation policy that, on the tile grid, acts as an adaptive local kernel operating on text rather than learned embeddings. This policy is an inspectable prompt that can be refined from model-observed tissue phenotypes via textual gradients, enabling automatic semantic optimization from local cellular context to broader tissue morphology without fine-tuning MLLM weights. On representative HER2 and CAMELYON16 regions, SLMP improves tile-level tumor description accuracy in settings spanning general-purpose and pathology-specialized backbones, with gains of +3.3 to +19.6 percentage points. Random-neighbor ablations confirm that these gains stem from spatial context rather than additional text alone, and inspecting the optimized policies reveals interpretable, tissue-specific decision rules. Besides, without any weight updates or fine-tuning the backbone MLLM, SLMP substantially improves general-purpose MLLMs and narrows its gap to pathology-specialized counterparts, offering a transparent and flexible mechanism for incorporating spatial reasoning into MLLM-based pathology analysis.

In social insects, mortality risk comprises distinct components that may not share the same predictors: lifespan duration, senescence trajectory, and thermal vulnerability. We tested these three axes in 18 Australian ant species using paired field-laboratory survival assays (2,363 cohort-day observations; 1,148 workers). Body size predicted duration (Cox HR = 0.67, p = 0.002), while colony size (p = 0.60) and the size x temperature interaction (p = 0.72) showed no detectable moderating effect. A weak but significant size x foraging-rate interaction was detected (LRT p = 0.014), suggesting that intrinsic physiology remains the most parsimonious explanation for the main size-longevity pattern, although ecological context may contribute. Senescence trajectory was associated with circadian niche rather than size: it was steepest in matinal species (Kruskal-Wallis p = 0.009; matinal vs. crepuscular p = 0.002) and was uncorrelated with body mass (Spearman p = 0.32). Thermal hazard plateaued above 20 degrees C (Delta AIC = -38; p < 0.001), with elevated thermal sensitivity in Rhytidoponera (Ectatomminae) above the plateau (5% per degree C, p = 0.015). Circadian regime and lineage identity, not body size, therefore emerge as the most climate-relevant axes, although they are strongly collinear (Cramer's V = 0.85). These results show that body size captures only one dimension of mortality risk and that size-based vulnerability indices may misrank taxa when senescence and thermal sensitivity are decoupled from body size.

Multi-strain epidemiological systems frequently exhibit self-sustained oscillations, yet severe algebraic complexity has long obstructed a complete analytical characterization of these dynamics. Seeking to bypass these barriers, we have identified a broad, analytically tractable class of two-strain models featuring asymmetric cross-immunity that precludes secondary infections for a single strain. This targeted structural simplification allows us to derive explicit expressions for coexistence equilibria and map their stability. We show that the relative transmission advantage of secondary infections fundamentally governs the onset of robust limit-cycle oscillations, extending previous narrow-borderline results. Crucially, numerical simulations reveal a rich macroscopic landscape where small-amplitude local oscillations coexist with large-amplitude, recurrent outbreak cycles. The structural robustness of these phenomena is confirmed by incorporating effects such as waning immunity and isolation and extending the framework beyond the analytically tractable class. These findings provide a clear mechanistic explanation for multi-strain epidemic cycles and offer a highly tractable baseline for future theoretical developments.

Self-organized spatial patterns are central to nonequilibrium physics and cell biology, yet locating instabilities in multi-component, reaction-diffusion networks remains challenging because standard eigenvalue analyses scale with the number of biochemical states and rely on reaction kinetics often poorly constrained by experiments. Exploiting the common mass-conserving structure of protein reaction kinetics and the fact that nonlinear feedback is typically confined to membrane reactions while lateral membrane diffusion is negligible, we develop a geometric classification that predicts stationary, and approximately also oscillatory, pattern-forming instabilities from reactive equilibria. The developed criteria reduce the stability analysis from the full component space to the space of conserved species. On this reduced space, slope matrices—describing the change of equilibrium cytosolic densities with respect to total species densities—govern onset. Based on densities in chemical equilibrium, this approach eliminates the requirement of comprehensive kinetic knowledge frequently lacking in experimental systems. Thus, a broad range of instabilities can be understood as mass-redistribution instabilities—self-amplifying mass redistribution caused by shifting local equilibria. We apply these criteria to models for the Escherichia coli Min system and the Caenorhabditis elegans polarity system and show that the reduction extends to the mixed-dimensional dynamics in systems coupling bulk cytosolic dynamics with membrane dynamics on the boundary. Together, these results provide an interpretable, broadly applicable, and experimentally accessible framework for diagnosing and designing pattern formation in multicomponent nonequilibrium systems on the basis of conservation laws.

Neurodegenexrative diseases such as Alzheimer's disease and Parkinson's disease are diagnosed most reliably only after substantial, often irreversible, neuronal loss has already occurred, creating an urgent need for quantitative tools that can detect subtle, early, and individual-specific brain changes from neuroimaging data. This review surveys a broad and rapidly evolving toolkit of data-driven techniques for translational neuroscience and personalized neuro-health, organized around four complementary methodological pillars. Throughout, we emphasize how these methodologically diverse approaches converge on a common translational goal: personalized, mechanistically grounded, and clinically actionable models of individual brain health, and we close by discussing the principal open statistical, computational, and clinical challenges that remain.

Environmental sharing changes the value of diversification even when the marginal statistics experienced by each lineage remain unchanged. A minimal model of cell division couples this effect to the inheritance of a conserved protective reserve. Each mother partitions its reserve between two daughters, and each fixed partition policy generates a random demographic operator whose top Lyapunov exponent determines long-term growth. Weak environmental sharing favors symmetric inheritance, whereas sufficiently shared innovation selects a separated asymmetric branch near {\alpha \simeq 0.2}. The transition therefore occurs by a finite branch crossing rather than by a continuous departure from equal partition. The asymmetric phase persists when the protection law or reserve turnover is changed, although its boundary depends on protection nonlinearity and reserve memory. Because shared and private environmental innovations have identical marginal statistics, the mean reproductive operator is independent of the shared-innovation fraction. A dominant-mode second-order approximation then separates the asymmetric advantage into a mean-operator growth cost of specialization and a reduction of sensitivity to collective fluctuations. This approximation predicts the finite asymmetric branch selected by the full random-operator dynamics, while the full operator product determines the numerical crossing. Environmental sharing can therefore drive symmetry breaking in the inheritance of a conserved protective resource when the loss of diversification between lineages increases the value of diversification generated at division.

Local chemical perception and property reasoning are both essential for understanding how molecular structure determines properties. Current LLM-based chemical reasoning methods either receive SMILES/molecular images together with descriptions of local motifs, or reason directly from molecular images. Neither approach enables the model to focus on chemically meaningful regions before reasoning. To address this gap, we propose Visual Latent Structural Reasoning (VLSR), an end-to-end framework that jointly learns localization and reasoning from molecular images. Central to our approach is a localize-then-reason strategy. VLSR first learns to locate chemically meaningful regions in a molecular image. It then reasons about their property effects in a compact latent workspace before producing the final answer. Under the same inference setup, this design achieves 9.6X higher throughput than a comparable textual-reasoning baseline.

Fast pixelated detectors are driving 4D-STEM toward microsecond dwell times, a regime in which the finite response of the scan deflection coils becomes comparable to the dwell time itself. Using direct probe imaging and sub-frame diffraction analysis, we document a significant intra-dwell scan-coil delay that systematically smears the recorded signal anisotropically along the fast scan direction, with a settling timescale of several tens of microseconds. We present a phase-correlation-based sub-frame alignment procedure that measures and corrects this smearing, and we assess its impact on focused and defocused 4D-STEM reconstructions over a range of scan step sizes. The correction restores signal across a broad range of spatial frequencies, with the largest gains at the large step sizes required for low-dose biological imaging. Because it operates on existing data with no modification to the microscope, the method offers a practical route to recovering signal that would otherwise be lost to scan-coil delay.

The field of bioinformatics struggles with legacy code - old code that is commonly used but may no longer have a maintainer, or may be written in an now-unfamiliar language (e.g. Perl, Fortran). This incurs maintenance cost (technical debt), but dynamically typed languages also negatively impacts the environment and fail to make use of modern hardware. Legacy code may also have security or safety problems that make it unsuited for use in clinical settings. Here we show that agentic AI, combined with static analysis, can be used to translate legacy code to the modern language Rust. We provide prompts and supporting software to aid systematic translation, and evaluate it on common software for NGS and imaging. We showcase the result on our software Bascet: Size was reduced by ~80x, build time decreased by ~10x, and performance of key steps improved >3x. Unix dependencies were also removed, making Bascet the only single-cell pipeline able to run on native Windows, without a container. Large-scale refactoring of bioinformatics software is thus now possible at a limited budget, enabling more complex tools to be developed.

Highly pathogenic avian influenza (HPAI) outbreaks require rapid control during active transmission and evidence-based decisions on the safe restocking of depopulated farms. We developed a stochastic spatial SEIR-based metapopulation model for a synthetic HPAI outbreak on the fictional Jolly Island. Farms were classified as `Broiler-2', `organic duck', or `Other' production systems. The model incorporated local, environmental, movement-mediated, and distance-dependent transmission, together with reactive and preventive culling, production-specific confinement, and capacity-based restocking. The simulated epidemic was geographically concentrated and differed substantially among production classes. Preventive culling reduced mean cumulative burden from 16,362.7 to 13,631.9 infectious-farm-days, with an overall reduction of 16.7%. Earlier confinement substantially reduced epidemic magnitude, while stronger environmental transmission increased the epidemic peak. Restocking risk declined as the epidemic approached resolution. Under the model assumptions, 24 May 2026 was the first candidate date satisfying the predefined rebound-probability threshold of 0.20. For restocking on 15 March 2026, none of the tested restocking fractions met this criterion. Capacity-based restocking reduced cumulative burden by 8.45% and rebound probability from 0.780 to 0.533, compared with restocking relative to the baseline population. These findings demonstrate the value of integrating epidemic control and post-outbreak recovery within a single modelling framework. Timely confinement, targeted preventive culling, and phased capacity-based restocking may reduce both epidemic burden and resurgence risk, although operational decisions should also incorporate surveillance, biosecurity, economic considerations, and regulatory requirements.

Neurodegenerative diseases can be viewed as spreading processes on brain networks, in which pathological proteins propagate between anatomically connected brain regions. Mathematical models have been used to study this process, but they generally ignore the influence of neuronal activity, even though experimental studies show that neuronal firing promotes protein transmission. Here, we couple a general node-activity process to susceptible–infected–susceptible dynamics. In this framework, an epidemic threshold determines whether small pathological seeds can grow, while a dominant network mode determines where growth begins. We derive approximations showing how neuronal activity shifts this threshold and redirects spreading by mixing structural network modes. For networks with multiscale structure, we decompose these changes into contributions from regional mean activity and within-region activity variation, allowing us to account for activity heterogeneity that is not resolved by brain imaging. Stochastic simulations validate the theoretical results across synthetic networks. We next use longitudinal human positron emission tomography to test whether neuronal activity predicts where and how broadly pathology spreads. Regional glucose metabolism serves as a proxy for neuronal activity, while tau accumulation measures disease progression. Adding neuronal activity to the network model captures spatial patterns of disease progression that are not explained by structural connectivity and established disease markers alone. Across individuals, predicted epidemic thresholds are also associated with how broadly pathology spreads through the brain. Together, these results connect epidemic theory to neurodegeneration, implicate neuronal activity as a driver of Alzheimer's disease progression, and motivate activity-modulating therapies to slow or prevent pathological spread.

Microbial populations generally evolve in fluctuating environments under time-varying conditions. These are often described by binary switching models, sometimes seen as coarse-grained feast-famine cycles, in which resource availability switches abruptly between abundant and scarce conditions. However, experimental studies suggest that feast-famine environments actually exhibit more complex temporal dynamics. Here, we study how two strains, one growing slightly slower than the other, compete for the same resources in fluctuating environments comprising a finite number of intermediate states, each having its own carrying capacity. Environmental switching between these states and their carrying capacities represents gradual changes in nutrient availability. This class of multi-state stochastic switching models can be interpreted as a coarse-grained description of feast–famine cycles and allows us to investigate strain competition under the gradual recovery and depletion of resources. By computational and analytical means, we characterise the population dynamics in these multi-state fluctuating environments. In particular, we study how the switching rates and distribution of carrying capacities affect the population-size statistics, fixation probability, and mean fixation time. By comparing these results with their counterparts in binary environments, we clarify how the frequency and amplitude of environmental fluctuations influence population dynamics in coarse-grained feast-famine cycles.

Biochemical reaction networks have become a central theoretical framework for implementing molecular computation. A key challenge is finite time computational accuracy, as computation outputs are encoded in limiting steady states (LSSs) of species concentrations while practical implementations operate for only finite time. This work proposes a concise characterization of convergence rate for biochemical computational modules with multiple output species, and rigorously establishes it as being bounded by the eigenvalue with largest (least negative) real part of the Jacobian matrix. Two numerical examples illustrate how the theoretical lower bound shapes the convergence rate range and reveals its dependence on reaction rate constants. This formulation enables systematic evaluation of biochemical computation speed and provides a practical design measure for constructing high-accuracy and error-controlled biochemical computational modules.

Protein language models (PLMs) have transferred the latest advances from natural language processing to computational biology. These models, trained on large corpora of protein sequence data, are widely used to translate amino acid sequences into latent-space embeddings, ready for use in diverse downstream tasks (DTs). By consensus, embeddings from the models' last layers are used, while the models' internal behavior remains poorly understood. We analyzed 13 PLMs across 15 DTs and 9 datasets to assess the value of embeddings from intermediate PLM layers. We trained probe models on embeddings from each layer, compared their performance, and showed that the last layers of PLMs rarely produced embeddings that led to the best results on downstream tasks. Furthermore, we identified a connection between how models learn a certain DT and the similarity between that DT and the pre-training objective. For example, for residue-level downstream tasks, we observed a steady increase in performance across almost all PLM layers, which we attributed to their similarity to most PLMs' pre-training objectives. To allow the community to capitalize on our findings, we provide PLMSommelier, a Python package that automatically identifies the best PLM layer for a given DT with ~98% accuracy and creates a truncated model using only the early layers up to the best-performing layer. This will help users save time and memory during inference and yield better predictive performance.

A predictive model can fit its data even when its information set is insufficient; fit alone cannot establish sufficiency. This Perspective introduces Level II-A, a new design-based inference framework to test this distinction, illustrated in anticipatory EEG using contingent negative variation. A pre-event endpoint is committed before the delay to the imperative event is randomised. That later-assigned delay thereby becomes a negative-control probe of whether past-adapted information was sufficient for an already committed result. Under the past-adapted factorisation, accounts using only pre-commitment information cannot systematically order the endpoint by that delay. Leakage-safe preprocessing, a frozen label-blind comparator and retained-sample qualifications carry the exclusion to the confirmatory residual. A qualified material negative ordering supports conditional insufficiency without identifying a mechanism; an adequately sensitive null supports a bounded affirmative conclusion calibrated by the pipeline's false-adequacy rate. A non-compensatory rule separates these from diagnostic failure, selection-limited, opposite-direction and inconclusive outcomes. No human EEG data are analysed. In the synthetic benchmark, grid-based false-adequacy boundaries are $15\,\mu\mathrm{V\,s^{-1}}$ for assignment isolation and $30\,\mu\mathrm{V\,s^{-1}}$ for the sequential e-value route, in both directions. The design transfers wherever endpoint commitment precedes an exogenous label, probing sufficiency only where a declared alternative predicts ordering by it. It turns "the past explains it" from a working explanatory assumption into a magnitude-qualified, testable claim.

The classical problem of Other Minds has dogged philosophers for millennia; asking if we have any way to truly understand the experience of another mind. We know our own intrinsic experience by acquaintance, but can only ever hope to possess an extrinsic description of another's, with the two separated by an acquaintance gap. Structural approaches aim to characterise experience in terms of a mathematical structure, and promise a 'Rosetta Stone'; that is, a principled method to translate the contents of experience into a mathematical structure. In this chapter, we examine two such approaches - the Qualia Structure Paradigm (Qstr) and Integrated Information Theory (IIT) - to ask what, if anything, they allow us to infer about another mind should they possess the Rosetta Stone they seek. Qstr proceeds inter-phenomenally; aiming to exhaustively characterise an experience by its relations to all other experiences and providing a necessary condition on the sameness of experience. IIT proceeds intra-phenomenally; conjecturing that a single experience is accounted for by the cause-effect structure unfolded from a substrate in a state, an explanatory identity that is both necessary and sufficient for the sameness of experience. While neither approach can cross the acquaintance gap to another's mind, they provide a common formal medium through which minds may be compared and in turn reduce the size of this gap. We identify the strength of inference with levels of structural correspondence in Category Theory. These descend from isomorphism through strong and weak adjunction, to a principled limit where inference runs out. We conclude that these structural approaches and the existence of their respective Rosetta Stones would not solve the problem of Other Minds, but instead provide a system of principled constraints on what we may infer, which is far more than what has been justifiable before.

Identifying individual learning styles optimizes pedagogical efficacy. While traditional questionnaires are structured, behavioral tracking methods require prolonged interaction log accumulation. To overcome these temporal constraints, this paper proposes an objective Electroencephalography (EEG) approach evaluating Phase Locking Value (PLV) connectivity against localized features across the Active-Reflective (AR) and Verbal-Visual (VV) Felder-Silverman dimensions. EEG signals were recorded from 28 participants during Raven's Advanced Progressive Matrices tasks. Support Vector Machine classification used Leave-One-Subject-Out Cross-Validation (LOSO-CV) alongside a 70:30 intra-subject split. The VV dimension achieved 70.00% subject-level accuracy driven by distinct fronto-occipital polarization. Conversely, the AR dimension yielded lower cross-subject generalizability (55.56%) due to overlapping executive networks and a "Systematic Neural Inversion" phenomenon, where stable individual connectivity signatures operated diametrically opposed to global boundaries (up to 20-0 voting margins). Ultimately, these outcomes demonstrate that rigid "one-size-fits-all" classifiers are bounded by biological diversity, emphasizing the need for future adaptive feature transformation techniques to bridge the cross-subject generalization gap.

Scaling relations linking species size, abundance, and resource availability are among the most robust empirical regularities in ecology. However, a mechanistic explanation for how these community-level laws emerge from ecological processes remains elusive. Here, we address this gap by developing a minimal spatially explicit dynamical framework for forest communities that incorporates seed dispersal, growth limited by local light availability, local competition, and global resource constraints grounded in metabolic scaling principles. By deriving an analytical solution for the tree-size distribution, we show that its stationary state exhibits two distinct power-law regimes whose exponents are controlled by the relative strength of resource and spatial competition. The crossover between these regimes is set by the interplay between seed injection and local resource availability, establishing an explicit link between the scaling exponent of the size distribution and forest condition. Finally, we show that boundary disturbances can break the ecological balance between competing species and induce effects that propagate deeply into the forest bulk, far beyond the single-plant dispersal range. Together, these results provide a unifying dynamical perspective on forest scaling laws with potential applications to a broad range of biological communities.

SVPLEX is a Nextflow pipeline for cohort-level structural variant detection from short-read whole-genome sequencing data. The pipeline implements six different structural variant callers with different strengths and weaknesses, integrating different levels of evidence for SVs, and generates a merged consensus callset across the analysis cohort. Callset filtering is achieved by leveraging consensus among multiple individual callers and by ensuring that deletion and duplication calls are supported by observable changes in read depth. The output merged cohort SV callset can then be used to assess cohort-specific variation, remove technical artefacts, and serve as input for rare disease variant prioritisation workflows. SVPLEX is user-friendly, reproducible, scalable, and can be executed flexibly on either a local workstation, a high-performance compute (HPC) cluster, or deployed on cloud infrastructure. The required inputs are alignment files for the cohort of interest, and the output is a single merged cohort structural variant VCF. SVPLEX is available on GitHub (bahlolab/SVPLEX) and is licensed under the MIT open-source licence.

Wastewater contains rich biological signals that can be used to monitor population health, track infectious diseases, and detect emerging outbreaks. Pathogens and other biomarkers in sewage provide a unique, noninvasive view of disease prevalence at the community level. However, extracting these signals requires extensive field sampling, laboratory analysis, and expert interpretation. Consequently, wastewater-based epidemiology (WBE) datasets are scarce, geographically fragmented, and rarely released as open data. Even when available, existing datasets typically cover short time periods and limited geographic regions, restricting their usefulness for method development, benchmarking, and large-scale modeling. To address this gap, we present an application of an existing patterns-of-life simulation framework for generating synthetic infectious disease and wastewater pathogen datasets for wastewater-based epidemiology. Our approach extends a patterns-of-life simulation framework by using it as a model of human mobility and behavior while incorporating disease transmission and pathogen shedding dynamics. The resulting framework generates high-resolution spatial and temporal datasets capturing infection dynamics, mobility behavior, and wastewater-associated pathogen signals. We release a fully simulated dataset containing check-in records, social network links, infection states, pathogen loads, and ground-truth disease transmission information. These data support controlled experimentation for outbreak detection, source localization, resource allocation, surveillance strategy design, mobility-aware wastewater analysis, and targeted public health interventions. We provide datasets for Fulton County and demonstrate that the framework generalizes to other regions by generating data for any city with available OpenStreetMap information.

Studies of biologically plausible learning rules (feedback alignment, predictive coding, STDP) train small networks on $32\times32$ images and compare them with brain responses to naturalistic stimuli presented at much higher resolution. We find that a common result in this setting, that untrained or locally trained networks rival or beat backpropagation at early visual cortex, depends on the resolution at which networks are evaluated. In representational similarity analysis (RSA) against human V1, the gap between an untrained and a backprop-trained network grows monotonically from $-0.001$ (95% CI $[-0.011, 0.009]$) at the 32 px training resolution to $+0.030$ $[0.018, 0.042]$ at 224 px (six resolutions, 5 seeds; per-subject RSA on cross-run stimulus pairs). The growth reproduces for three further learning rules, directionally in single-seed macaque electrophysiology, along training, and for an ImageNet ResNet-50 and a Swin-Tiny transformer, which also align best at low resolution although trained at 224 px. Gabor and pixel structure, the normalization state of the untrained baseline, and a global brightness statistic do not account for it. Against a criterion fixed before the analysis, limiting image content to 32 px while the input still grows does not shrink the V1 gap ($+0.035$ $[0.023, 0.046]$ vs. $+0.030$ at 224 px; ratio 1.14 $[1.03, 1.34]$): it arises once the input exceeds the training resolution, through a decline of backprop. A single luminance value per image reaches $\rho = 0.054$ against V1, matching the untrained network (0.053), which bounds what RSA on globally pooled features can resolve at V1 in this dataset. The one learning effect that holds across resolution is backprop above untrained at LOC. Comparisons of learning rules or architectures at early visual cortex need to control, and report, the evaluation resolution.

Congruency effects, observed in conflict tasks such as Stroop and flanker tasks, have been investigated for nearly a century in psychology and neuroscience, but their mechanistic basis is not fully understood. We introduce a verbal-only LLM conflict task in which a prompt stem elicits a default same-color completion and an explicit rule either agrees with (congruent condition) or conflicts with (incongruent condition) the completion. Gemma-2-2B and six Pythia models ranging from 410M to 12B parameters showed strong default same-color tendencies, and six of seven models showed strong congruency effects. Using causal attribution analysis, attention analysis, and attention ablations, we identified distinct processing pathways in these LLMs: a pathway involving short-range attention to a superficial color cue that is preferentially activated in the congruent condition, and a pathway involving long-range attention to the rule prefix that is preferentially activated in the incongruent condition. Fine-tuning that strengthened the default same-color tendency had divergent effects on task conditions, reducing incongruent performance while increasing congruent performance. In contrast, increasing rule set size selectively impaired incongruent performance. These converging findings support an account in which congruency effects in this task arise from competition between an in-weight default mapping and an in-context rule-based mapping. More broadly, our findings illustrate how LLMs can serve as model systems for mechanistic analysis of competition between default and rule-governed response tendencies within a single learned network.

Hit-to-lead optimization requires iterative design of hit analogs across competing potency, selectivity, physicochemical, pharmacokinetic, safety, and synthetic constraints. We present SABLE (Synthetically-accessible Agentic Bayesian Ligand Exploration), an open-source framework that employs natural-language orchestration to guide chemical structure optimization. SABLE uses an LLM to interpret user-defined goals and route tasks, while specialized tools perform reaction-templated analog enumeration, physicochemical and ADMET property prediction, structure-based affinity scoring, and Bayesian optimization. The resulting workflow is a computational twin of the analytical and prioritization stages of the design-make-test-analyze cycle, providing provenance of each numerical output. Across single, and multi-objective optimization studies, SABLE enriches candidate sets for user-defined computational objectives while evaluating only a subset of the enumerated search space. Its modular architecture allows tools and characterization backends to be replaced by editing a simple config file, without modifying operational logic. SABLE provides an extensible decision-support framework for prioritizing synthetically constrained analogs in early-stage drug discovery.

AlphaFold3-class structure predictors pair a representational trunk, which processes sequence and context, with a diffusion module, which generates atomic coordinates. How biological information changes as it crosses this architectural boundary remains poorly understood. We analyze per-residue activations from the Pairformer trunk and diffusion module of Boltz-1 using linear probes, sparse autoencoders (SAEs), and causal interventions. From the trunk, both geometry (secondary structure, disorder) and sequence chemistry (amino-acid identity, signal peptides, disulfide-bond annotations) are linearly decodable. In the diffusion module, the two diverge. Secondary structure transfers essentially unchanged, whereas sequence chemistry is strongly attenuated. We then test whether decodable directions can steer the model, intervening on the final trunk single representation that conditions the diffusion module. Helix and coil directions change predicted structure dose-dependently against matched-norm random controls, but a beta-strand direction that is highly predictive (F1 =0.82) produces no measurable increase in strand content: linear decodability does not imply causal influence at the site we tested. The same probes also score markedly lower against sparse SwissProt annotations than against dense DSSP labels, because unannotated residues that the model gets right are charged as false positives; such scores are therefore lower bounds. Finally, supervised probes outscore single SAE features wherever a label already exists. We release the trained trunk and diffusion SAEs, Boltz-1 per-residue activations, and the analysis code.

Drug response prediction (DRP) models are an active area of research in pharmacogenomics, with growing potential to accelerate the identification of effective anticancer drugs. However, their predictive performance is often constrained by limited dataset scale and insufficient coverages of cancer and chemical spaces. In addition, inconsistent benchmarking practices hinder reliable comparison across models. Standardized frameworks, such as the Innovative Methodologies and New Data for Predictive Oncology Model Evaluation (IMPROVE) project, provide unified data schemas and evaluation protocols for consistent benchmarking, but improving model generalizability requires larger and more diverse training data. In this work, we substantially expand the IMPROVE benchmark through large-scale integration of pharmacogenomic data, primarily from PharmacoDB, together with additional smaller data sources. The expanded resource includes millions of drug response measurements, broader multi-omics coverage, and a major increase in chemical diversity, adding more than 50,000 compounds. To evaluate the impact of the new dataset compared to the original IMPROVE benchmark dataset, we trained DRP models using the two datasets and assess their prediction performance using a common test set and several evaluation strategies, including drug-blind, cancer-blind, and disjoint data splits. While cancer-blind performance remained comparable to the original benchmark, models trained on the expanded dataset showed consistent improvements in drug-blind and disjoint settings, indicating enhanced generalization to previously unseen compounds. These results position the expanded dataset as a community resource that provides a richer foundation for developing DRP models intended to aid in the discovery of novel anticancer drugs.