arXiv Quantitative Biology @arxiv_bio@qoto.org

Termites which are able to forage in the open can be often seen, in the field or in the lab: (i) wandering around, forming no observable pattern, or (ii) clustering themselves in a dense and almost immobile pack, or (iii) milling about in a circular movement. Despite been well reported patterns, they are normally regarded as independent phenomena whose specific traits have never been properly quantified. Evidence, however, favours the hypothesis that these are interdependent patterns, arisen from self-organised interactions and movement among workers. After all, termites are a form of active matter where blind cooperative individuals are self-propelled and lack the possibility of visual cues to spatially orientate and align. It follows that their non-trivial close-contact patterns could generate motion-collision induced phase separations. This would then trigger the emergence of these three patterns (disorder, clustering, milling) as parts of the same continuum. By inspecting termite groups confined in arenas, we could quantitatively describe each one of these patterns in detail. We identified disorder, clustering and milling spatial patterns. These phases and their transitions are characterised aiming to offer refinements in the understanding of these aspects of self-propelled particles in active matter where close-range contacts and collisions are important.

Accurate blood glucose prediction can enable novel interventions for type 1 diabetes treatment, including personalized insulin and dietary adjustments. Although recent advances in transformer-based architectures have demonstrated the power of attention mechanisms in complex multivariate time series prediction, their potential for blood glucose (BG) prediction remains underexplored. We present a comparative analysis of transformer models for multi-horizon BG prediction, examining forecasts up to 4 hours and input history up to 1 week. The publicly available DCLP3 dataset (n=112) was split (80%-10%-10%) for training, validation, and testing, and the OhioT1DM dataset (n=12) served as an external test set. We trained networks with point-wise, patch-wise, series-wise, and hybrid embeddings, using CGM, insulin, and meal data. For short-term blood glucose prediction, Crossformer, a patch-wise transformer architecture, achieved a superior 30-minute prediction of RMSE (15.6 mg / dL on OhioT1DM). For longer-term predictions (1h, 2h, and 4h), PatchTST, another path-wise transformer, prevailed with the lowest RMSE (24.6 mg/dL, 36.1 mg/dL, and 46.5 mg/dL on OhioT1DM). In general, models that used tokenization through patches demonstrated improved accuracy with larger input sizes, with the best results obtained with a one-week history. These findings highlight the promise of transformer-based architectures for BG prediction by capturing and leveraging seasonal patterns in multivariate time-series data to improve accuracy.

Social perception unfolds as we freely interact with people around us. We investigated the neural basis of real world face perception using multi electrode intracranial recordings in humans during spontaneous interactions with friends, family, and others. Computational models reconstructed the faces participants looked at during natural interactions, including facial expressions and motion, from brain activity alone. The results highlighted a critical role for the social vision pathway, a network of areas spanning parietal, temporal, and occipital cortex. This network was more sharply tuned to subtle expressions compared to intense expressions, which was confirmed with controlled psychophysical experiments. These findings reveal that the human social vision pathway encodes facial expressions and motion as deviations from a neutral expression prototype during natural social interactions in real life.

Cell type annotation is a critical yet laborious step in single-cell RNA sequencing analysis. We present a trustworthy large language model (LLM)-agent, CellTypeAgent, which integrates LLMs with verification from relevant databases. CellTypeAgent achieves higher accuracy than existing methods while mitigating hallucinations. We evaluated CellTypeAgent across nine real datasets involving 303 cell types from 36 tissues. This combined approach holds promise for more efficient and reliable cell type annotation.

This article introduces a computational method, called "Recapture of Diffusive Agents & Particle Swarm Optimization" (RDA-PSO), designed to estimate the dispersal parameter of diffusive insects in mark-release-recapture (MRR) field experiments. In addition to describing the method, its properties are discussed, with particular focus on robustness in estimating the observed diffusion coefficient. It is shown that RDA-PSO provides a simple and reliable approach to quantify mosquito dispersal that outperforms other techniques based on the solution of the diffusion equation, three of which are also introduced in this work. Examples of application to both synthetic and real field data for the yellow fever mosquito are provided.

Peptides are recognized for their varied self-assembly behaviors, forming a wide array of structures and geometries, such as spheres, fibers, and hydrogels, each presenting a unique set of material properties. The functionalities of these materials hold exceptional interest for applications in biology, medicine, photonics, nanotechnology and the food industry. In specific, the ability to exploit peptides as viable and sustainable mechanical materials requires sequence design that enables superior performance, notably a high Young's modulus. As the peptide sequence space is vast, however, even a slight increase in sequence length leads to an exponential increase in the number of potential peptide sequences to be characterized. Here, we combine coarse-grained molecular dynamics simulations, atomic force microscopy experiments and machine learning models to correlate the sequence length and composition with the mechanical properties of self-assembled peptides. We calculate the Young's modulus for all possible amino acid sequences of di- and tripeptides using high-throughput coarse-grained methods, and validate these calculations through in-situ mechanical characterization. For pentapeptides, we select and calculate properties for a subset of sequences to train a machine learning model, which allows us to predict the modulus for other sequences. The combined workflow not only identifies promising peptide candidates with exceptional mechanical performances, but also extends current understanding of the sequence-to-function relationships for peptide materials, for specific applications.

The emergence of telomere-to-telomere (T2T) genome assemblies has opened new avenues for comparative genomics, yet effective tokenization strategies for genomic sequences remain underexplored. In this pilot study, we apply Byte Pair Encoding (BPE) to nine T2T primate genomes including three human assemblies by training independent BPE tokenizers with a fixed vocabulary of 512,000 tokens using our custom tool, dnaBPE. Our analysis reveals that only 11,569 tokens are shared across all assemblies, while nearly 991,854 tokens are unique to a single genome, indicating a rapid decline in shared vocabulary with increasing assembly comparisons. Moreover, phylogenetic trees derived from token overlap failed to recapitulate established primate relationships, a discrepancy attributed to the disproportionate influence of species-specific high-copy repetitive elements. These findings underscore the dual nature of BPE tokenization: while it effectively compresses repetitive sequences, its sensitivity to high-copy elements limits its utility as a universal tool for comparative genomics. We discuss potential hybrid strategies and repeat-masking approaches to refine genomic tokenization, emphasizing the need for domain-specific adaptations in the development of large-scale genomic language models. The dnaBPE tool used in this study is open-source and available at https://github.com/aglabx/dnaBPE.

Ultradian rhythms - quasi-rhythmic fluctuations in behavior and physiology with periods shorter than 24 hours - are observed across various organisms, including humans. Despite their role in key biological processes such as sleep architecture and hormone regulation, their underlying mechanisms remain poorly understood. Here, we leveraged wearable sensor technology for continuous monitoring of physiological signals in 16 healthy participants over two weeks. By systematically removing circadian and longer-scale rhythms, we isolated ultradian dynamics and modeled them using the Hankel Alternative View of Koopman (HAVOK) framework,a data-driven approach based on Takens' embedding theorem and Koopman operator theory. This allowed us to characterize ultradian rhythms as an intermittently forced linear system and distinguish between regular oscillatory behavior and more complex dynamics. Across participants, ultradian fluctuations were well-described by the HAVOK model, with intermittent forcing consistently observed. The model demonstrated strong forecasting accuracy, with root mean squared error (RMSE) of $0.0315 \pm 0.02$, $0.0306 \pm 0.02$, and $0.0218 \pm 0.02$ in the leading time-delay coordinates. Notably, a significant sex difference in model rank (z = -2.06, p = 0.0396) suggests that sex hormones may play a key role in ultradian dynamics. These findings provide evidence for intermittently forced linear systems as a useful framework for understanding ultradian rhythms and their regulation.

Biochemical systems are inherently stochastic, particularly those with small-molecule populations. The spatial distribution of molecules plays a critical role and requires the inclusion of spatial coordinates in their analysis. Stochastic models such as the chemical master equation are commonly used to study these systems. However, analytical solutions are limited to specific cases, and stochastic simulations require significant computational resources. To mitigate these challenges, approximation methods, such as the moment approach, reduce the system to a set of ordinary differential equations, thereby lowering the computational requirements. This study investigates the conditions under which the second-moment approach yields exact results during the dynamic evolution of chemical reaction-diffusion networks. The analysis encompasses second-order or higher-order reactions and Hill functions without relying on higher-order moment estimations or closure approximations. Starting with stationary states, we extended the analysis to a dynamic evolution. An enzymatic process and an antithetic feedback system were examined for purely reactive systems, demonstrating the approach's accuracy in capturing system behavior and quantifying errors. The study was further extended to genetic regulatory networks governed by Hill functions, including both purely reactive and reaction-diffusion systems, validating the method in spatially distributed contexts. This framework enables precise characterization of biochemical systems, avoiding information loss typically associated with approximations and allowing for stability analysis under fluctuations. These findings optimize computational strategies while providing insights into intracellular signaling and regulatory processes, paving the way for efficient and accurate stochastic modeling in biochemical systems.

Given usefulness of protein language models (LMs) in structure and functional inference, RNA LMs have received increased attentions in the last few years. However, these RNA models are often not compared against the same standard. Here, we divided RNA LMs into three classes (pretrained on multiple RNA types (especially noncoding RNAs), specific-purpose RNAs, and LMs that unify RNA with DNA or proteins or both) and compared 13 RNA LMs along with 3 DNA and 1 protein LMs as controls in zero-shot prediction of RNA secondary structure and functional classification. Results shows that the models doing well on secondary structure prediction often perform worse in function classification or vice versa, suggesting that more balanced unsupervised training is needed.

This study conducts a Quantitative Structure Property Relationship (QSPR) analysis to explore the correlation between the physical properties of drug molecules and their topological indices using machine learning techniques. While prior studies in drug design have focused on degree-based topological indices, this work analyzes a dataset of 166 drug molecules by computing degree-distance-based topological indices, incorporating vertex-edge weightings with respect to different six atomic properties (atomic number, atomic radius, atomic mass, density, electronegativity, ionization). Both linear models (Linear Regression, Lasso, and Ridge Regression) and nonlinear approaches (Random Forest, XGBoost, and Neural Networks) were employed to predict molecular properties. The results demonstrate the effectiveness of these indices in predicting specific physicochemical properties and underscore the practical relevance of computational methods in molecular property estimation. The study provides an innovative perspective on integrating topological indices with machine learning to enhance predictive accuracy, highlighting their potential application in drug discovery and development processes. This predictive may also explain that establishing a reliable relationship between topological indices and physical properties enables chemists to gain preliminary insights into molecular behavior before conducting experimental analyses, thereby optimizing resource utilization in cheminformatics research.

Recent studies have demonstrated the feasibility of modeling single-cell data as natural languages and the potential of leveraging powerful large language models (LLMs) for understanding cell biology. However, a comprehensive evaluation of LLMs' performance on language-driven single-cell analysis tasks still remains unexplored. Motivated by this challenge, we introduce CellVerse, a unified language-centric question-answering benchmark that integrates four types of single-cell multi-omics data and encompasses three hierarchical levels of single-cell analysis tasks: cell type annotation (cell-level), drug response prediction (drug-level), and perturbation analysis (gene-level). Going beyond this, we systematically evaluate the performance across 14 open-source and closed-source LLMs ranging from 160M to 671B on CellVerse. Remarkably, the experimental results reveal: (1) Existing specialist models (C2S-Pythia) fail to make reasonable decisions across all sub-tasks within CellVerse, while generalist models such as Qwen, Llama, GPT, and DeepSeek family models exhibit preliminary understanding capabilities within the realm of cell biology. (2) The performance of current LLMs falls short of expectations and has substantial room for improvement. Notably, in the widely studied drug response prediction task, none of the evaluated LLMs demonstrate significant performance improvement over random guessing. CellVerse offers the first large-scale empirical demonstration that significant challenges still remain in applying LLMs to cell biology. By introducing CellVerse, we lay the foundation for advancing cell biology through natural languages and hope this paradigm could facilitate next-generation single-cell analysis.

In this paper we present GeoThinneR, an R package for efficient and flexible spatial thinning of species occurrence data. Spatial thinning is a widely used preprocessing step in species distribution modeling (SDM) that can help reduce sampling bias, but existing R implementations rely on brute-force algorithms that scale poorly with large datasets. GeoThinneR implements multiple thinning approaches, including ensuring a minimum distance between points, subsampling points on a grid, and filtering based on decimal precision. To handle large datasets, it introduces two optimized algorithms based on local kd-trees and adaptive neighbor estimation, which greatly reduce memory usage and execution time. Additional functionalities such as group-wise thinning and point prioritization are included to facilitate its use in SDM workflows. We here provide performance benchmarks using both simulated and real-world data to demonstrate substantial performance improvements over existing tools.

New methods of CGM data analysis are emerging that are valuable for interpreting CGM patterns and underlying metabolic physiology. These new methods use functional data analysis and artificial intelligence (AI), including machine learning (ML). Compared to traditional metrics for evaluating CGM tracing results (CGM Data Analysis 1.0), these new methods, which we refer to as CGM Data Analysis 2.0, can provide a more detailed understanding of glucose fluctuations and trends and enable more personalized and effective diabetes management strategies once translated into practical clinical solutions.

Understanding cell identity and function through single-cell level sequencing data remains a key challenge in computational biology. We present a novel framework that leverages gene-specific textual annotations from the NCBI Gene database to generate biologically contextualized cell embeddings. For each cell in a single-cell RNA sequencing (scRNA-seq) dataset, we rank genes by expression level, retrieve their NCBI Gene descriptions, and transform these descriptions into vector embedding representations using large language models (LLMs). The models used include OpenAI text-embedding-ada-002, text-embedding-3-small, and text-embedding-3-large (Jan 2024), as well as domain-specific models BioBERT and SciBERT. Embeddings are computed via an expression-weighted average across the top N most highly expressed genes in each cell, providing a compact, semantically rich representation. This multimodal strategy bridges structured biological data with state-of-the-art language modeling, enabling more interpretable downstream applications such as cell-type clustering, cell vulnerability dissection, and trajectory inference.

A significant advancement in bioinformatics is using genome graph techniques to improve variation discovery across organisms. Traditional approaches, such as bwa mem, rely on linear reference genomes for genomic analyses but may introduce biases when applied to highly diverse bacterial genomes of the same species. Pangenome graphs provide an alternative paradigm for evaluating structural and minor variations within a graphical framework, including insertions, deletions, and single nucleotide polymorphisms. Pangenome graphs enhance the detection and interpretation of complex genetic variants by representing the full genetic diversity of a species. In this study, we present a robust and reliable bioinformatics pipeline utilising the PanGenome Graph Builder (PGGB) and the Variation Graph toolbox (vg giraffe) to align whole-genome sequencing data, call variants against a graph reference, and construct pangenomes from assembled genomes. Our results demonstrate that leveraging pangenome graphs over a single linear reference genome significantly improves mapping rates and variant calling accuracy for simulated and actual bacterial pathogens datasets.

Protein structure generative models have seen a recent surge of interest, but meaningfully evaluating them computationally is an active area of research. While current metrics have driven useful progress, they do not capture how well models sample the design space represented by the training data. We argue for a protein Frechet Inception Distance (FID) metric to supplement current evaluations with a measure of distributional similarity in a semantically meaningful latent space. Our FID behaves desirably under protein structure perturbations and correctly recapitulates similarities between protein samples: it correlates with optimal transport distances and recovers FoldSeek clusters and the CATH hierarchy. Evaluating current protein structure generative models with FID shows that they fall short of modeling the distribution of PDB proteins.

Superspreading events are known to disproportionally contribute to onwards transmission of epidemic and pandemic viruses. Preventing infections at a small number of high-transmission settings is therefore an attractive public health goal. Here, we use deterministic and stochastic mathematical modelling to quantify the impact of intranasal sprays in containing outbreaks at a known superspreading event (the 2020 SARS-CoV-2 outbreak at the Diamond Princess cruise ship) and a conference event that led to extensive transmission. We find that in the Diamond Princess cruise ship case study, there exists a 7-14-day window of opportunity for widespread prophylactic spray usage to significantly impact the number of infections averted. Given an immediate response to a known SARS-CoV-2 outbreak, alongside testing and social distancing measures, prophylactic efficacy and coverage greater than 65% could reduce the average number of infections by over 90%. In the conference case study, in the absence of additional public health interventions, analyses suggest much higher prophylactic efficacies and coverages are required to achieve a similar outcome. However, prophylactic use can half an individual's probability of being infected, and significantly reduce the probability of developing a severe infection. These results suggest that at a known potential superspreading event, early use of intranasal sprays can complement quarantining measures and significantly suppress a SARS-CoV-2 outbreak, even at suboptimal coverage. At a potential superspreading event of short duration, intranasal sprays can reduce individuals' risk of infection, but cannot prevent all infections or onwards community transmission.

Spiking neural networks (SNNs) are powerful mathematical models that integrate the biological details of neural systems, but their complexity often makes them computationally expensive and analytically untractable. The firing rate of an SNN is a crucial first-order statistic to characterize network activity. However, estimating firing rates analytically from even simplified SNN models is challenging due to 1) the intricate dependence between the nonlinear network dynamics and parameters, and 2) the singularity and irreversibility of spikes. In this Letter, we propose a class of computationally efficient, simulation-free estimators of firing rates. This is based on a hierarchy of Markovian approximations that reduces the complexity of SNN dynamics. We show that while considering firing rates alone is insufficient for accurate estimations of themselves, the information of spiking synchrony dramatically improves the estimator's accuracy. This approach provides a practical tool for brain modelers, directly mapping biological parameters to firing rate.

We employ a unified framework for computing the information capacity of biological signaling systems using Fisher Information. By deriving closed-form or easily computable information capacity formulas, we quantify how well different signaling models, including binomial, multinomial, Poisson, Gaussian, and Gamma distributions, can discriminate among input signals. These expressions clarify how key features such as signal range, noise scaling, pathway length, and receivers' diversity shape the theoretical limits of sensing. In particular, we show how signal-to-noise ratio and fold-change sensitivity arise naturally within the Fisher formalism, and how signal degradation in cascades imposes linear information loss. Our results provide intuitive, analytically grounded tools to benchmark and guide the analysis of real signaling systems, without requiring computationally expensive mutual information estimation. While motivated by cellular communication, the framework generalizes to any system where noisy input-output relationships constrain transmission fidelity, including synthetic biology, sensor networks, and engineered communication channels.