Vegetation in semi-arid ecosystems frequently organizes into spatially heterogeneous mosaics that regulate ecosystem functioning, productivity, and resilience. These patterns arise from local biological interactions, including facilitation among neighboring plants and competition for limiting resources. Classical theoretical approaches have attributed such organization to scale-dependent feedbacks, predicting regular spatial patterns and abrupt transitions to collapse. However, growing empirical and theoretical evidence reveal that environmental variability and demographic stochasticity can fundamentally reshape spatial organization, driving irregular clusters, dynamic mosaics, and gradual rather than catastrophic vegetation declines. In drylands, rainfall variability is a dominant source of environmental forcing: precipitation typically occurs in short, irregular pulses that transiently enhance survival and recruitment before competitive interactions again dominate. Near persistence thresholds, ecosystem dynamics are therefore governed not only by average climatic conditions but also by the timing and spatial coincidence of favorable events. Under these conditions, positive density dependence and local facilitation can critically determine whether vegetation patches persist, expand, or collapse. Here, we develop an individual-based model that integrates intermittent precipitation with local Allee effects to examine how stochastic rainfall shapes spatial organization and persistence. We show that the interaction between pulsed resource availability and density-dependent survival generates irregular cluster structures and strongly modulates extinction risk, with resilience emerging from local spatial covariance and neighborhood density rather than from total biomass alone. These results highlight the importance of individual-level, stochastic processes in determining ecosystem resilience.

Coinfection phenomena are common in nature, yet there is a lack of analytical approaches for coinfection systems with a high number of circulating and interacting strains. In this paper, we investigated a coinfection SIS framework applied to N strains, co-circulating in a structured host population. Adopting a general formulation for fixed host classes, defined by arbitrary epidemiological traits such as class-specific transmission rates, susceptibilities, clearance rates, etc., our model can be easily applied in different frameworks: for example, when different host species share the same pathogen, in classes of vaccinated or non-vaccinated hosts, or even in classes of hosts defined by the number of contacts. Using the strain similarity assumption, we identify the fast and slow variables of the epidemiological dynamics on the host population, linking neutral and non-neutral strain dynamics, and deriving a global replicator equation. This global replicator equation allows to explicitly predict coexistence dynamics from mutual invasibility coefficients among strains. The derived global pairwise invasion fitness matrix contains explicit traces of the underlying host population structure, and of its entanglement with the strain interaction and trait landscape. Our work thus enables a more comprehensive study and efficient simulation of multi-strain dynamics in endemic ecosystems, paving the way to deeper understanding of global persistence and selection forces, jointly shaped by pathogen and host diversity.

The mechanical response of biological soft tissues is influenced by wall heterogeneity, including spatial variations in wall thickness. Traditional models for homogeneous soft tissues under uniaxial loading predict higher stretch and stress in thinner regions. In fact, large gradients in stretch and stress are predicted to be induced by spatial variations in wall thickness. In prior studies, the role of collagen fibers in regions of thickness transition has been largely neglected or only considered in terms of their effect on anisotropy. Here, we explore the role of collagen fibers as primary load-bearing components across regions of varying wall thickness, using a three-dimensional representative volume element (RVE) model incorporating explicit collagen fiber architecture and a gradual thickness gradient. We examined two distinct collagen fiber configurations across the thickness transition: one featuring abrupt fiber termination and another with fiber continuity. Finite element analysis (FEA) under uniaxial tension revealed that load transfer by continuous fibers across the specimen markedly reduced the importance of the change in wall thickness, with stretch differentials dropping from ~20% (fiber-termination network) to 0.68% (continuous fibers) and stress differentials dropping from ~65% (fiber-termination network) to 2.3% (continuous fibers). Fiber tortuosity delayed the point at which mechanical response was governed by fiber structure. These findings demonstrate the critical role of fiber continuity in reducing stretch and stress gradients across regions of varying wall thickness and clarify the importance of accurately representing fiber architecture when modeling soft tissues with heterogeneous wall thickness.

Neurons are spatially extended cells; different parts of a neuron have specific voltage dynamics. Important types of neurons even generate different spikes in different parts of the cell. Neurons' inputs are also often spatially compartmentalized, with different sources targeting different locations on the cell. Classic mean-field theories for neural population activity, however, rely on point-neuron models with at most one type of spike. Here, we develop a statistical field-theoretic approach to understanding collective activity in networks of compartmental neurons, including those generating multiple types of spikes. We use this to examine simple models of networks with thick-tufted layer 5 pyramidal cells, which generate calcium spikes in their apical dendrite when dendritic depolarization coincides with a back-propagating somatic spike. In the weakly-coupled regime, we uncover an exact mean-field limit for these networks that maps them to a marked point process. We use this mean-field limit to compare the impact of compartmentalized recurrent excitatory and inhibitory connectivity on the equilibrium phase diagram. This exposes regions of metastability between various activity states, including activity with silent vs active dendrites, with and without inhibitory activity, and oscillations.

Recent advancements in digital pathology have enabled comprehensive analysis of Whole-Slide Images (WSI) from tissue samples, leveraging high-resolution microscopy and computational capabilities. Despite this progress, there is a lack of labeled datasets and open source pipelines specifically tailored for analysis of skin tissue. Here we propose Histo-Miner, a deep learning-based pipeline for analysis of skin WSIs and generate two datasets with labeled nuclei and tumor regions. We develop our pipeline for the analysis of patient samples of cutaneous squamous cell carcinoma (cSCC), a frequent non-melanoma skin cancer. Utilizing the two datasets, comprising 47,392 annotated cell nuclei and 144 tumor-segmented WSIs respectively, both from cSCC patients, Histo-Miner employs convolutional neural networks and vision transformers for nucleus segmentation and classification as well as tumor region segmentation. Performance of trained models positively compares to state of the art with multi-class Panoptic Quality (mPQ) of 0.569 for nucleus segmentation, macro-averaged F1 of 0.832 for nucleus classification and mean Intersection over Union (mIoU) of 0.907 for tumor region segmentation. From these predictions we generate a compact feature vector summarizing tissue morphology and cellular interactions, which can be used for various downstream tasks. Here, we use Histo-Miner to predict cSCC patient response to immunotherapy based on pre-treatment WSIs from 45 patients. Histo-Miner identifies percentages of lymphocytes, the granulocyte to lymphocyte ratio in tumor vicinity and the distances between granulocytes and plasma cells in tumors as predictive features for therapy response. This highlights the applicability of Histo-Miner to clinically relevant scenarios, providing direct interpretation of the classification and insights into the underlying biology.

Alzheimer's disease (AD) exhibits substantial clinical and biological heterogeneity, complicating efforts in treatment and intervention development. While new computational methods offer insights into AD progression, the reproducibility of these subtypes across datasets remains understudied, particularly concerning the robustness of subtype definitions when validated on diverse databases. This study evaluates the consistency of AD progression subtypes identified by the Subtype and Stage Inference (SuStaIn) algorithm using T1-weighted MRI data across 5,444 subjects from ANMerge, OASIS, and ADNI datasets, forming four independent cohorts. Each cohort was analyzed under two conditions: one using the full cohort, including cognitively normal controls, and another excluding controls to test subtype robustness. Results confirm the three primary atrophy subtypes identified in earlier studies: Typical, Cortical, and Subcortical, as well as the emergence of rare and atypical AD variants such as posterior cortical atrophy (PCA). Notably, each subtype displayed varying robustness to the inclusion of controls, with certain subtypes, like Subcortical, more influenced by cohort composition. This investigation underscores SuStaIn's reliability for defining stable AD subtypes and suggests its utility in clinical stratification for trials and diagnosis. However, our findings also highlight the need for improved dataset diversity, particularly in terms of ethnic representation, to enhance generalizability and support broader clinical application.

INTRODUCTION: Previous studies have applied normative modeling on a single neuroimaging modality to investigate Alzheimer Disease (AD) heterogeneity. We employed a deep learning-based multimodal normative framework to analyze individual-level variation across ATN (amyloid-tau-neurodegeneration) imaging biomarkers. METHODS: We selected cross-sectional discovery (n = 665) and replication cohorts (n = 430) with available T1-weighted MRI, amyloid and tau PET. Normative modeling estimated individual-level abnormal deviations in amyloid-positive individuals compared to amyloid-negative controls. Regional abnormality patterns were mapped at different clinical group levels to assess intra-group heterogeneity. An individual-level disease severity index (DSI) was calculated using both the spatial extent and magnitude of abnormal deviations across ATN. RESULTS: Greater intra-group heterogeneity in ATN abnormality patterns was observed in more severe clinical stages of AD. Higher DSI was associated with worse cognitive function and increased risk of disease progression. DISCUSSION: Subject-specific abnormality maps across ATN reveal the heterogeneous impact of AD on the brain.

A wide range of applications and research has been done with genome-scale metabolic models. In this work we describe a methodology for comparing metabolic networks constructed from genome-scale metabolic models and how to apply this comparison in order to infer evolutionary distances between different organisms. Our methodology allows a quantification of the metabolic differences between different species from a broad range of families and even kingdoms. This quantification is then applied in order to reconstruct phylogenetic trees for sets of various organisms.

Background: Nowadays, the reconstruction of genome scale metabolic models is a non-automatized and interactive process based on decision taking. This lengthy process usually requires a full year of one person's work in order to satisfactory collect, analyze and validate the list of all metabolic reactions present in a specific organism. In order to write this list, one manually has to go through a huge amount of genomic, metabolomic and physiological information. Currently, there is no optimal algorithm that allows one to automatically go through all this information and generate the models taking into account probabilistic criteria of unicity and completeness that a biologist would consider. Results: This work presents the automation of a methodology for the reconstruction of genome scale metabolic models for any organism. The methodology that follows is the automatized version of the steps implemented manually for the reconstruction of the genome scale metabolic model of a photosynthetic organism, {\it Synechocystis sp. PCC6803}. The steps for the reconstruction are implemented in a computational platform (COPABI) that generates the models from the probabilistic algorithms that have been developed. Conclusions: For validation of the developed algorithm robustness, the metabolic models of several organisms generated by the platform have been studied together with published models that have been manually curated. Network properties of the models like connectivity and average shortest mean path of the different models have been compared and analyzed.

In this contribution, a design of a synthetic calibration genetic circuit to characterize the relative strength of different sensing promoters is proposed and its specifications and performance are analyzed via an effective mathematical model. Our calibrator device possesses certain novel and useful features like modularity (and thus the possibility of being used in many different biological contexts), simplicity, being based on a single cell, high sensitivity and fast response. To uncover the critical model parameters and the corresponding parameter domain at which the calibrator performance will be optimal, a sensitivity analysis of the model parameters was carried out over a given range of sensing protein concentrations (acting as input). Our analysis suggests that the half saturation constants for repression, sensing and difference in binding cooperativity (Hill coefficients) for repression are the key to the performance of the proposed device. They furthermore are determinant for the sensing speed of the device, showing that it is possible to produce detectable differences in the repression protein concentrations and in turn in the corresponding fluorescence in less than two hours. This analysis paves the way for the design, experimental construction and validation of a new family of functional genetic circuits for the purpose of calibrating promoters.

We introduce several axioms which may or may not hold for any given subgraph of the directed graph of all organisms (past, present and future) where edges represent biological parenthood, with the simplifying background assumption that life does not go extinct. We argue these axioms are plausible for species: if one were to define species based purely on genealogical relationships, it would be reasonable to define them in such a way as to satisfy these axioms. The main axiom we introduce, which we call the identical ancestor point axiom, states that for any organism in any species, either the species contains at most finitely many descendants of that organism, or else the species contains at most finitely many non-descendants of that organism. We show that this (together with a convexity axiom) reduces the subjectivity of species, in a technical sense. We call connected sets satisfying these two axioms "specieslike clusters." We consider the question of identifying a set of biologically plausible constraints that would guarantee every organism inhabits a maximal specieslike cluster subject to those constraints. We provide one such set consisting of two constraints and show that no proper subset thereof suffices.

The relationship between traits that influence pathogen virulence and transmission is part of the central canon of the evolution and ecology of infectious disease. However, identifying directional and mechanistic relationships among traits remains a key challenge in various subfields of biology, as models often assume static, fixed links between characteristics. Here, we introduce learning evolutionary trait relationships (LETR), a data-driven framework that applies Granger-causality principles to determine which traits drive others and how these relationships change over time. LETR integrates causal discovery with generative mapping and transfer-operator analysis to link short-term predictability with long-term trait distributions. Using a synthetic myxomatosis virus-host data set, we show that LETR reliably recovers known directional influences, such as virulence driving transmission. Applying the framework to global pandemic (SARS-CoV-2) data, we find that past virulence improves future transmission prediction, while the reverse effect is weak. Invariant-density estimates reveal a long-term trend toward low virulence and transmission, with bimodality in virulence suggesting ecological influences or host heterogeneity. In summary, this study provides a blueprint for learning the relationship between how harmful a pathogen is and how well it spreads, which is highly idiosyncratic and context-dependent. This finding undermines simplistic models and encourages the development of new theory for the constraints underlying pathogen evolution. Further, by uniting causal inference with dynamical modeling, the LETR framework offers a general approach for uncovering mechanistic trait linkages in complex biological systems of various kinds.

The tumor-immune system plays a critical role in colorectal cancer progression. Recent preclinical and clinical studies showed that combination therapy with anti-PD-L1 and cancer vaccines improved treatment response. In this study, we developed a multiscale mathematical model of interactions among tumors, immune cells, and cytokines to investigate tumor evolutionary dynamics under different therapeutic strategies. Additionally, we established a computational framework based on approximate Bayesian computation to generate virtual tumor samples and capture inter-individual heterogeneity in treatment response. The results demonstrated that a multiple low-dose regimen significantly reduced advanced tumor burden compared to baseline treatment in anti-PD-L1 therapy. In contrast, the maximum dose therapy yielded superior tumor growth control in cancer vaccine therapy. Furthermore, cytotoxic T cells were identified as a consistent predictive biomarker both before and after treatment initiation. Notably, the cytotoxic T cells-to-regulatory T cells ratio specifically served as a robust pre-treatment predictive biomarker, offering potential clinical utility for patient stratification and therapy personalization.

In this study, we develop a metapopulation model framework to identify optimal harvesting strategies for a population in a stream network. We consider two distinct optimization objectives: maximization of total biomass and maximization of total yield, under the constraint of a fixed total harvesting effort. We examine in detail the special case of a two-patch network and fully characterize the optimal strategies for each objective. We show that when the population growth rate exceeds a critical threshold, a single harvesting strategy can simultaneously maximize both objectives. For general $n$-patch networks with homogeneous growth rates across patches, we focus on the regime of large growth rates and demonstrate that the optimal harvesting strategy selects patches according to their intraspecific competition rates and an effective net flow metric determined by network connectivity parameters.

Agent-based cellular models simulate tissue evolution by capturing the behavior of individual cells, their interactions with neighboring cells, and their responses to the surrounding microenvironment. An important challenge in the field is scaling cellular resolution models to real-scale tumor simulations, which is critical for the development of digital twin models of diseases and requires the use of High-Performance Computing (HPC) since every time step involves trillions of operations. We hereby present a scalable HPC solution for the molecular diffusion modeling using an efficient implementation of state-of-the-art Finite Volume Method (FVM) frameworks. The paper systematically evaluates a novel scalable Biological Finite Volume Method (BioFVM) library and presents an extensive performance analysis of the available solutions. Results shows that our HPC proposal reach almost 200x speedup and up to 36% reduction in memory usage over the current state-of-the-art solutions, paving the way to efficiently compute the next generation of biological problems.

Virtual cell modeling aims to predict cellular responses to diverse perturbations but faces challenges from biological complexity, multimodal data heterogeneity, and the need for interdisciplinary expertise. We introduce CellForge, a multi-agent framework that autonomously designs and synthesizes neural network architectures tailored to specific single-cell datasets and perturbation tasks. Given raw multi-omics data and task descriptions, CellForge discovers candidate architectures through collaborative reasoning among specialized agents, then generates executable implementations. Our core contribution is the framework itself: showing that multi-agent collaboration mechanisms - rather than manual human design or single-LLM prompting - can autonomously produce executable, high-quality computational methods. This approach goes beyond conventional hyperparameter tuning by enabling entirely new architectural components such as trajectory-aware encoders and perturbation diffusion modules to emerge from agentic deliberation. We evaluate CellForge on six datasets spanning gene knockouts, drug treatments, and cytokine stimulations across multiple modalities (scRNA-seq, scATAC-seq, CITE-seq). The results demonstrate that the models generated by CellForge are highly competitive with established baselines, while revealing systematic patterns of architectural innovation. CellForge highlights the scientific value of multi-agent frameworks: collaboration among specialized agents enables genuine methodological innovation and executable solutions that single agents or human experts cannot achieve. This represents a paradigm shift toward autonomous scientific method development in computational biology. Code is available at https://github.com/gersteinlab/CellForge.

A major goal of neuroscience is to understand brain computations during visual processing in naturalistic settings. A dominant approach is to use image-computable deep neural networks trained with different task objectives as a basis for linear encoding models. However, in addition to requiring estimation of a large number of linear encoding parameters, this approach ignores the structure of the feature maps both in the brain and the models. Recently proposed alternatives factor the linear mapping into separate sets of spatial and feature weights, thus finding static receptive fields for units, which is appropriate only for early visual areas. In this work, we employ the attention mechanism used in the transformer architecture to study how retinotopic visual features can be dynamically routed to category-selective areas in high-level visual processing. We show that this computational motif is significantly more powerful than alternative methods in predicting brain activity during natural scene viewing, across different feature basis models and modalities. We also show that this approach is inherently more interpretable as the attention-routing signals for different high-level categorical areas can be easily visualized for any input image. Given its high performance at predicting brain responses to novel images, the model deserves consideration as a candidate mechanistic model of how visual information from retinotopic maps is routed in the human brain based on the relevance of the input content to different category-selective regions.