Polygenic risk scores (PRS) developed in European populations often show reduced predictive performance in non-European populations, limiting their clinical utility. This lack of transferability across ancestries remains a major challenge in genomic medicine and raises concerns about health equity. We aimed to evaluate whether uncertainty-aware prediction, implemented through Mondrian Cross-Conformal Prediction, improves the performance and reliability of polygenic risk score-based predictions across ancestries for nephropathy, stroke, and myocardial infarction in individuals with type 2 diabetes in a multi-ethnic cohort. We leveraged Mondrian Cross-Conformal Prediction (MCCP), an uncertainty quantification framework, combined with logistic regression applied to a multi-polygenic risk score (multiPRS) to predict the risk of nephropathy, stroke, and myocardial infarction in individuals with type 2 diabetes. Two training frameworks were evaluated: one using 4,098 individuals with type 2 diabetes of European ancestry from the ADVANCE trial for training and 17,574 White British, 1,145 South Asian, and 749 African UK Biobank participants for testing; and another using the 17,574 White British UK Biobank participants for training and the South Asian and African participants for testing. Logistic regression provided robust baseline performance across populations. On top of this baseline, MCCP did not improve performance but added capabilities absent from probability-based stratification: for each individual, it issued a prediction together with an explicit confidence and credibility level; it allowed a tolerated error level to be set in advance and delivered prediction sets respecting it in the majority of settings; and it flagged individuals for whom no reliable prediction could be made. Applying MCCP to PRS-based prediction thus enables uncertainty-aware risk stratification and improves the reliability of risk prediction across ancestries, providing a more equitable framework for clinical use.

The average mechanical power output (AMPO) over a stretch-shortening cycle produced by a muscle depends on muscle length and stimulation over time. While the effects of cycle frequency and muscle length excursion on AMPO are well-known, several questions remain about the effects of muscle length and stimulation over time on the maximally attainable AMPO. For example, which precise muscle length and stimulation over time yield maximal AMPO? In situ experiments are inherently limited to a finite set of muscle length and stimulation over time. To overcome this limitation, we combined in situ experiments on rat m. gastrocnemius medialis with Hill-type muscle modelling. We first performed dedicated trials to estimate the muscle-tendon-complex (MTC) properties of each rat. Subsequently, we performed various stretch-shortening cycles with substantial differences in cycle frequency, shortening-to-lengthening time ratio and MTC length excursion. Model-predicted AMPO correlated nearly perfectly with experimentally measured AMPO (r[2] > 0.98). This justified further exploration using the Hill-type MTC model. Using the Hill-type MTC model, we predicted that AMPO peaks at a cycle frequency of 3.5 Hz, with a shortening-to-lengthening time ratio of 6:1 and an MTC length excursion of 8 mm. Notably, cycle frequency and MTC length excursion showed a strong interaction: increasing one necessitated a decrease in the other to maximise AMPO. By contrast, the optimal shortening-to-lengthening time ratio remained remarkably constant across all tested combinations of cycle frequency and MTC length excursion. This shows that muscles should spend substantially more time shortening than lengthening to maximise AMPO.

REvolutionH-tl is a fast, scalable, and integrated software platform for inferring orthology relationships, gene trees, species trees, and reconciled evolutionary scenarios directly from sequence data. Built upon the formal framework of best match graphs (BMGs), REvolutionH-tl predicts orthogroups and orthologous gene pairs with high accuracy, requiring neither precomputed trees nor multiple external tools. The software reconstructs event-labeled gene and species trees, seamlessly integrating reconciliation to produce fast, accurate, and biologically insightful evolutionary scenarios. Through extensive benchmarking on synthetic datasets with known ground truth, REvolutionH-tl outperforms or matches the accuracy of established tools such as OrthoFinder, Proteinortho, RAxML, GeneRax, and RANGER-DTL, while achieving significantly lower runtimes. A key innovation of REvolutionH-tl is its built-in support for detailed, publication-ready visualizations, which allow users to explore genome evolution dynamics, orthogroup composition, and reconciliation results with clarity and ease. These visual features position REvolutionH-tl as the first platform of its kind to combine analytical precision with intuitive interpretability. The software is open-source, cross-platform, and freely available at https://pypi.org/project/revolutionhtl/, providing a robust solution for large-scale evolutionary analyses in comparative genomics.

In vitro colonies of the unicellular parasite Trypanosoma brucei expand radially and establish fingering instabilities, a collective behavior known as social motility. The underlying mechanisms are thought to involve single-cell motility, chemical communication among cells, and mechanical interactions with the liquid boundary, but their relative contributions remain unclear. We aimed to determine which of the mechanisms are necessary to quantitatively reproduce the morphological characteristics of social motility. We developed a two-dimensional agent-based model that simulates colonies of 105-106 cells at single-cell resolution-two to four orders of magnitude larger than previous models. Cells are represented as point particles executing directional random walks with auto-chemotactic alignment and exponential colony growth. The colony boundary is modelled using a grid-based approach in which interactions with agents can locally weaken and expand it. The model was quantitatively evaluated by applying our previously established morphology metrics. We show quantitative agreement of the simulation results and experimental data in terms of colony morphology. Parameter exploration revealed that finger formation arises within a narrow range of trypanosome motility parameters that balance stochasticity and alignment, while boundary conditions modulate the speed of colony expansion. The diffusion coefficient of the chemotactic signal is the key determinant of pattern formation. Realistic behavior occurs at 2×10-11 - 10-10 m2/s which corresponds to molecules of 12.1-1690 kDa. These results demonstrate that complex colony morphologies can emerge from minimal cell-level rules, suggesting testable hypotheses for the molecular drivers of trypanosome social motility. Furthermore, our approach provides a framework for dissecting the interplay between motility, signaling, and mechanical confinement in other microbial systems exhibiting collective behavior.

Epidemic forecasting research often assesses ensembles and their component models using probabilistic scoring rules. Quantifying how individual models affect ensemble performance is challenging, particularly across multiple targets and spatial scales.

Deep learning is transforming the study of animal sound, enabling the automated identification of species, individuals, behaviors, and ecological patterns from large collections of recordings. While bioacoustic machine-learning models are growing more powerful, many biologists-ecologists, behavioral scientists, conservationists-and others working with acoustic data feel unprepared to navigate the computational workflows required to implement them. This article presents practical guidelines covering the full lifecycle of bioacoustic machine learning, including problem definition, data sourcing and annotation, model training, evaluation, deployment, reproducibility, and ethical considerations. Rather than providing a linear checklist, the guidelines outline an iterative framework for building science-led workflows, leveraging transfer learning and open-source tools, evaluating models based on the real-world cost of errors, and addressing domain shift under variable field conditions. Ultimately, this workflow demystifies the software-engineering process, providing a low-barrier and reproducible pathway for researchers applying machine learning to animal sound.

Ancient DNA provides a direct window into the evolutionary processes that have shaped living microbial species today, as well as their now extinct relatives. Advances in both sequencing methods and de novo assembly techniques have not only resulted in a flood of modern metagenomic sequencing data, but they have also allowed palaeogenomicists to retrieve vast amounts of ancient DNA from past microorganisms, including species and strains without modern reference genomes. However, the degraded nature of ancient DNA means that the standard techniques of genome assembly developed for modern DNA are unlikely to perform effectively, unless heavily modified. This hinders the incorporation of ancient data into broader metagenomic studies that would otherwise benefit from having deep time information on the evolution of different microbial species. In this primer and protocol paper, we provide guidance on ways to adapt existing metagenomic de novo assembly processes, including data input, tools, and settings, in order to perform more robustly and effectively on ancient DNA. After assembly, we then further describe how ancient DNA contigs can be identified and validated. The key steps of ancient metagenomic assembly are now integrated in a dedicated ancient DNA mode in the established pipeline nf-core/mag. By introducing support for ancient DNA data in nf-core/mag, we aim to improve the ability of researchers to more regularly integrate de novo assembled ancient microbial data into broader metagenomics studies of microbial ecology and evolution.

In recent years, high dimensional omics analyses have become more commonplace for investigating complex biological systems. Typically, these studies attempt to identify key biomolecules associated with a particular biological process. Often, machine learning (ML) is used to identify these biomolecules, typically by learning which biomolecules are highly predictive of a treatment, biological outcome, or phenotype. A major challenge of applying ML to high throughput omics is overcoming noise when sample size is limited and unbalanced with respect to tens of thousands of biomolecules measured. Thus, feature selection (the process of reducing the number of predictors) is both a critical and common step in the ML analysis pipeline. While much attention has been given to embedding and wrapping techniques for feature selection in the omics space, filter-based methods for model-free feature selection have appealing theoretical properties. This manuscript evaluates sure screening, a class of filter-based feature selection methods which provide analytical guarantees for true feature set retention. Here, we cover existing feature screening methods based on the sure screening principal, available software, methods to improve feature screening, and contextualize feature screening in the larger discussion of feature selection for omics data analysis. Additionally, a suite of model-free sure screening approaches is applied and compared for several omics biomedical applications in a ML classification context. We identified BcorSIS as the most effective and computationally efficient screening method across various omics datasets, consistently outperforming others like CSIS and DCSIS in runtime.

Cells make hard calls under noise. When signaling is abnormal, those calls can go wrong and drive pathological conditions and diseases. In this research, we develop a Neyman-Pearson (NP) detection-theory framework that maximizes probability of detection (PD) for a chosen false alarm probability (PFA), without requiring prior probabilities, using experimental single-cell measurements of NF-κB responses to tumor necrosis factor (TNF), a critical pathway involved in cell survival, apoptosis, immune signaling, and stress response, in wild-type and A20-deficient fibroblasts. We model log-responses as (multi)variate Gaussian and compute optimal thresholds, PD-PFA trade-offs, and ROC curves at 30 minutes and 4 hours. The NP framework captures expected biology: PD increases with TNF dose; wild-type cells outperform A20-/- at matched conditions; and combining two time points (bivariate analysis) improves detection (e.g., for 0.0052 vs. 0.2 ng/mL, PD rises from 0.71 (30 minutes) and 0.42 (4 hours) to 0.80 at PFA = 0.1). The analysis recovers expected biology (higher TNF causes higher detectability; negative feedback lowers late responses) and flags cases where decision quality degrades (e.g., perturbations that blunt separation between conditions). The same recipe extends to multivariate readouts without changing the logic. A non-parametric (kernel-density) detector applied to the raw single-cell data reproduces these detection probabilities (|ΔPD| ≤ 0.092), confirming the conclusions do not rely on the Gaussian approximation. Overall, the NP detection framework provides a compact, quantitative score of pathway performance and failure. It turns noisy single-cell readouts into actionable decision metrics that compare doses, time points, and perturbations, and ultimately, can help explain when and how cellular decisions drift toward pathology.

The origin and organizing principles of the genetic code remain central problems in molecular evolution. The low probability of the natural codon-to-amino acid mapping arising by chance has spurred the hypothesis that its structure is optimized for robustness to mutations and translational errors. For the construction of effective molecular machines, the repertoire of encoded amino acids must also be diverse enough in physicochemical features. Here, we examine whether the standard genetic code can be understood as a near-optimal solution balancing these two objectives: minimizing error load and aligning codon assignments with the naturally occurring amino acid composition. Using simulated annealing, we explore this trade-off across a broad range of parameters. We find that the standard genetic code resides near an optimum in the fitness landscape of possible genetic codes. The degeneracy of the code plays a dual role, minimizing mistranslation errors while matching codon multiplicity to amino acid usage frequencies. As a result, uniform codon usage alone is sufficient to recover the empirical amino acid composition, without any additional bias. It is a highly effective solution that balances fidelity against resource availability constraints. A comparative analysis of natural variants also reveals a functional decoupling: error robustness acts as a rigid global constraint determined by code topology, whereas compositional alignment serves as a more flexible variable that adapts to lineage-specific demands. These results support a multi-objective optimization framework in which the genetic code reflects a balance between translational fidelity and proteomic demand.

Analytically deriving equilibria, which often govern the long-term behavior of biochemical reaction networks, is essential for understanding cellular decision-making and robustness, yet remains computationally challenging for large and complex systems. The COMPILES framework of Hernandez et al. addresses this problem via network decomposition but is limited by a computationally intensive translation step and by its restriction to networks with zero kinetic deficiency. We introduce CRITERIA (Computing paRametrized posITive EquilibRIA), a new framework that overcomes these limitations through two key advances. First, it replaces the translation mechanism of COMPILES with a more efficient graph-theoretic formulation based on the work of Johnston and Burton, which generates a reaction-to-reaction graph from elementary flux modes and identifies directed cycles of the constructed graph via a binary linear program. Second, CRITERIA computes equilibria on a single unified translated network rather than solving subnetworks independently, thereby eliminating interdependencies that previously required extensive symbolic manipulation. Across a benchmark set of 26 biochemical models, CRITERIA achieves consistent and often substantial speed improvements while expanding applicability beyond the restricted class of zero kinetic deficiency systems. We demonstrate the biological utility of the framework by analyzing multistationarity, which underlies cellular decision-making, and absolute concentration robustness, a key mechanism for maintaining stable biochemical outputs, in the EnvZ-OmpR signaling pathway and a large-scale CRISPRi toggle switch, respectively. By improving both scalability and applicability, CRITERIA enables a systematic equilibrium analysis in biochemical networks of realistic size and complexity, providing a practical tool for studying long-term dynamical behavior in systems biology.

Since its introduction in 1968, Influenza A (H3N2) has undergone continuous antigenic evolution, necessitating frequent vaccine updates. To predict antigenicity and characterize antigenic drift without multiple sequence alignments, we present FluEmbed, a computational framework that leverages protein language models. FluEmbed accurately quantified the antigenic impact of viral evolution from RNA sequences, achieving strong predictive performance against hemagglutination inhibition (HI) assay titers (Spearman correlation: ρ = 0.67-0.80). FluEmbed also outperformed sequence-distance baselines (e.g., Hamming and BLOSUM62) and phylogenetic tree-based models that require sequence alignment. Using this model, we conducted in-silico mutagenesis experiments to identify site/amino acid combinations that differentially impacted antigenicity. To systematically investigate how specific mutations influence immune escape, we defined two classes of mutations: 'constrained', where only the most likely amino acid changes at historically mutation-prone sites were considered (thereby limiting the mutation space) and 'unconstrained', where all possible substitutions were allowed, providing a full exploration of potential antigenic shifts. Constrained mutations often confer limited antigenic changes, whereas unconstrained mutations exhibit greater escape potential, particularly outside the dominant viral lineages. Notably, 3C.2a was the only major lineage in which constrained and unconstrained mutations showed no significant difference (p ≈ 0.9), suggesting ongoing intra-clade competition rather than inter-lineage antigenic replacement. By enabling rapid, alignment-free antigenic prediction directly from sequence data, FluEmbed could complement traditional HI assays in real-time influenza surveillance and inform vaccine strain selection decisions.

High-throughput experimental platforms now routinely generate data from dozens or hundreds of independent observations. Simulation-based inference (SBI) offers a powerful framework for estimating model parameters from such complex datasets, but standard methods struggle to scale to the noisy multiple-replicates regime without incurring prohibitive computational costs or careful hyperparameter tuning. Here, we introduce a new method for fast and robust collective posterior inference from multiple independent replicates using a robust product-of-experts aggregation scheme that automatically mitigates the influence of outliers. Evaluating it on synthetic and empirical evolutionary datasets, we find it achieves state-of-the-art estimation accuracy and computational efficiency, including inference from noisy observations. Our method is compatible with any SBI framework, providing a scalable, plug-and-play solution for inference from noisy multiple-replicate datasets.

[This corrects the article DOI: 10.1371/journal.pcbi.1014133.].

[This corrects the article DOI: 10.1371/journal.pcbi.1014092.].