Convergent evolution, or the independent acquisition of similar phenotypes in distinct lineages, provides a powerful framework for investigating genomic changes associated with a phenotype. This paper details an update to RERconverge, a powerful R package that tests for associations between gene relative evolutionary rates (RERs) and convergent phenotypes to infer genomic regions associated with traits or selective pressures. We introduce new customizable analysis choices and scalable and efficient algorithms that can process larger genomic datasets, a critical improvement as genomic data become available for more species. Modifications to core functions in the RERconverge pipeline resulted in an immense speedup (by a factor of up to 28.6). The function that tests for associations between phenotypes and RERs has been expanded to include two new analytical methods for outlier control; we also provide here a summary of the statistical tests users can perform, along with their use cases. The code and walkthrough vignettes for the package are available at https://github.com/nclark-lab/RERconverge . Nathan L. Clark nclark@pitt.edu ; Maria Chikina mchikina@pitt.edu.
Simultaneous optimization of chromatographic resolution and analysis time constitutes a significant analytical challenge in multicomponent pharmaceutical analysis, as resolution-driven optimization strategies may improve peak separation without providing explicit control over retention time, often resulting in unnecessarily prolonged runtimes. To address this limitation, an optimization strategy based on a Box-Behnken experimental design was implemented in conjunction with the Improved Chromatographic Response Function (ICRF), which integrates separation quality and analysis time within a single mathematical objective function. This strategy describes the development of a rapid, chemometrically optimized UPLC-PDA method for the simultaneous determination of dorzolamide hydrochloride (DH) and timolol maleate (TI) in a commercial ophthalmic preparation. A Box-Behnken experimental design and optimization approach was employed in combination with an Improved Chromatographic Response Function (ICRF), which integrates resolution, peak overlap, peak width, and runtime into a single composite objective function. This strategy enabled short runtime (or short retention time of analytes in a chromatogram) while preserving adequate peak separation. Under the optimized conditions, complete chromatographic separation was achieved within 3 min using a BEH C18 column and a mobile phase consisting of acetonitrile and 4 × 10-4 M CCl3COOH (60:40, v/v) at a flow rate of 0.32 mL/min with detection at 275 nm. The method demonstrated excellent linearity over the range of 5.0-40.0 µg/mL (r > 0.999), with limits of detection of 0.51 µg/mL for DH and 0.61 µg/mL for TI. Mean recoveries were 99.9% and 99.5% for DH and TI, respectively, with satisfactory precision and robustness. The proposed ICRF-assisted optimization approach provided high-resolution separation within minimal runtime and was successfully applied to the routine quality-control analysis of a commercial ophthalmic formulation. The study demonstrates the effectiveness of composite response-based chemometric optimization in enhancing analytical efficiency in pharmaceutical drug analysis.
Performance bottlenecks in widely used genomics and bioinformatics software present a substantial and growing burden as biological datasets continue to increase in size and number. Relieving these bottlenecks relies largely on expert manual optimization and therefore remains difficult to scale. Here we present AutoZyme, an agentic framework for scientific software optimization. Given a target function, AutoZyme builds benchmarks, identifies bottlenecks, and iteratively tests code changes, retaining only those that improve runtime while preserving output. We evaluated AutoZyme on 45 functions, improving runtime without substantial memory increases in over 95% of cases considered. Across 38 functions from Seurat, Scanpy and related packages in genomics and bioinformatics, AutoZyme reduced runtime by a median of 8.52-fold, with the largest reductions exceeding 676-fold. The optimized functions are distributed through AutoZyme-Library as drop-in replacements for existing analysis pipelines. We also release AutoZyme as a reusable framework for optimizing additional user-specified packages and functions.
fractional-order modeling provides a powerful framework for representing memory-dependent conduction in excitable biological media. However, existing soliton-based models of myelinated nerve fibers are often theoretical, operator-specific, and insufficiently benchmarked in terms of numerical reproducibility, physiological plausibility, and computational cost. This study aims to compare the Liouville-Caputo, Atangana-Baleanu, and Beta fractional operators for modeling soliton-like action-potential propagation in ephaptically coupled myelinated nerve fibers, with emphasis on waveform stability, energy retention, biological consistency, computational efficiency, and adaptive parameter learning. A comparative computational modeling study was conducted using a coupled fractional nonlinear partial differential equation framework, physiological parameter mapping, numerical sensitivity analysis, and physics-informed neural network-based parameter estimation. A coupled fractional Korteweg-de Vries-type system was solved under identical initial and boundary conditions for the three fractional operators. The time-fractional order α was varied over [0.6, 1.0], while the space-fractional order β was varied over [1.5, 2.0]. Simulations used a uniform spatial grid, fixed time step, localized sech2 initial pulse, and Neumann boundary conditions. The operators were compared using soliton-like velocity, amplitude, pulse width, normalized energy retention, residual error, RMSE, MAE, and CPU runtime. A physics-informed neural network was further used to estimate model parameters while enforcing the fractional PDE residual. The Beta derivative produced the most localized and stable soliton-like pulses, with stronger amplitude preservation, lower energy loss, and shorter runtime than the Liouville-Caputo and Atangana-Baleanu formulations under the tested settings. Increasing ephaptic coupling strength reduced pulse amplitude, whereas increasing α improved propagation velocity and increasing β enhanced waveform localization. Quantitative residual and error analyses confirmed that the Beta-based formulation maintained low numerical error while preserving biologically plausible conduction behavior. The results support the Beta derivative as a biologically plausible and computationally efficient approximation for soliton-like nerve-pulse propagation in coupled myelinated fibers. The Liouville-Caputo and Atangana-Baleanu operators remain valuable for long-memory and fading-memory regimes, respectively. Future work should integrate literature-constrained biological consistency assessment, stochastic ion-channel dynamics, and heterogeneous multidimensional nerve-bundle geometries.
Long runtime, high memory demands, and reliance on high-performance computing increasingly limit the evolutionary analysis of long phylogenomic datasets. We review a scalable framework based on phylogenomic subsampling and upsampling (PSU), in which many small subsamples of sites from a long concatenated sequence alignment are extended by upsampling prior to inference, and the resulting analyses are then aggregated to obtain stable evolutionary estimates. PSU exploits a useful distinction between the computational burden and the inferential power of statistical methods in molecular phylogenetics: computational cost is strongly influenced by the number of distinct site patterns in the concatenated alignment, whereas statistical power depends primarily on the amount of evolutionary information represented by sites and substitutions. By reducing the former while restoring the latter through upsampling, PSU can approximate many full-data analyses at substantially lower computational cost. Evidence from simulated and empirical datasets shows that PSU can accurately estimate bootstrap support values, select optimal substitution models, test evolutionary hypotheses, and infer branch lengths, divergence times, and associated uncertainty measures, while often reducing runtime and memory requirements by orders of magnitude. The same subsampling-upsampling-aggregation principle underlies all of these applications. PSU also provides distributions of inferred clade support across independent subsamples, enabling detection of concordant and conflicting phylogenetic signals that may remain hidden in conventional concatenated phylogenomic analyses. Adaptive procedures for selecting the subsample size, the number of subsamples, and the number of upsampling replicates make the framework practical across diverse datasets. We suggest that PSU is a general strategy for scalable phylogenomic inference across a broad range of statistical methods. By enabling rigorous analyses of genome-scale alignments on standard computing hardware, PSU expands access to computationally intensive evolutionary methods while reducing the environmental and infrastructural costs of big-data phylogenomics.
In the field of stereo depth sensing, modern research predominantly prioritizes accuracy, yet inference speed remains a critical bottleneck for practical, real-time applications on resource-constrained platforms. Existing acceleration approaches often rely on lighter network architectures or runtime-specific optimizations, which may require architectural redesign, platform-specific tuning, or accuracy trade-offs. However, a common inefficiency remains in many stereo pipelines: feature extraction is typically performed using two separate forward passes, one for the left image and one for the right, even though both passes use the same network weights. We address this redundancy by concatenating the left and right images into a single combined tensor, enabling feature extraction in one batched pass while preserving the original network architecture. By reducing feature extraction time by up to 48.4%, our results demonstrate that this method accelerates the overall inference rate by 10% to 39% on average on Nvidia V100 and up to 28.4% on edge device, depending on the model architecture. This speedup is achieved at the expense of only a moderate increase in runtime memory consumption, while retaining the original accuracy. Because the method does not alter the core stereo network, it can be applied as a plug-and-play enhancement to both existing and newly developed stereo matching models.
Exact interpretable learning is attractive in regulated decision settings, but solver runtime can vary substantially across datasets and solver families. We introduce structural meta-features derived from Feature Interaction Graphs (FIG s) as interpretable signals for solver selection. We construct FIG s from binarized tabular data using pairwise mutual information and extract topology-aware signatures such as density and estimated treewidth. Using a transparent shallow decision-tree selector, we demonstrate that FIG features establish an interpretable structural view of solver behavior, complementing basic, statistical, and landmarking meta-features. Experiments on OpenML classification tasks show that topology-aware profiling exposes meaningful structural variation across datasets, although benchmark saturation prevents clear end-to-end routing gains over strong simple baselines. Our results validate FIG as a principled, interpretable diagnostic tool for algorithm selection in exact learning; its diagnostic relevance becomes apparent on harder instances where solver runtime separation is substantial.
Qualitative thematic analysis is widely used in health research to examine patient experiences and inform the refinement of digital health interventions, but it is time- and labor-intensive. Large language models (LLMs) may help accelerate this process, yet their performance may depend not only on the model itself but also on how the analytic workflow is structured. Current evidence remains limited on how different LLMs perform across multistage thematic analysis workflows and across multiple health-related qualitative datasets. This study aimed to evaluate a modular human-artificial intelligence (AI) collaboration pipeline for LLM-assisted thematic analysis and compare how model choice and workflow strategy influence alignment between AI-generated and human-generated themes across 3 qualitative health studies. The framework was applied to analyze deidentified semistructured interview transcripts from 3 completed qualitative health studies involving patients with interstitial lung disease, postural orthostatic tachycardia syndrome, and chronic obstructive pulmonary disease. Three LLMs were compared: Gemini (Gemini 3 Pro), ChatGPT (GPT-5.2-thinking), and Opus (version 4.6). The workflow separated analysis into code extraction, code combination, and theme generation, and 5 strategies were tested. AI-generated themes were embedded using sentence-t5-xxl and compared with human-generated themes using cosine similarity after alignment with Hungarian and Greedy matching. Runtime and output-format consistency were also examined. Output volume differed substantially by model. Gemini generated the fewest codes and themes, while ChatGPT showed a similar but higher output ceiling. Opus produced the largest and most variable codebooks and theme sets. Across the 3 studies, Opus showed the strongest and most consistent alignment with human-generated themes, with the best cosine similarity scores observed in postural orthostatic tachycardia syndrome-direct coding (mean 0.893, SD 0.041), chronic obstructive pulmonary disease-direct grouping (mean 0.891, SD 0.027), and interstitial lung disease-L3 (mean 0.889, SD 0.032). ChatGPT was competitive in selected settings, whereas Gemini generally produced slightly lower similarity scores but had the shortest runtime. ChatGPT and Opus also showed better formatting consistency and workflow usability than Gemini. A modular human-AI pipeline can support thematic analysis across multiple digital health interview studies, but performance depends strongly on both model choice and workflow design. Opus produced the most consistently human-aligned themes, while Gemini and ChatGPT showed different trade-offs in speed, fidelity, and usability. These findings support the use of LLMs as structured, human-supervised analytic assistants rather than replacements for qualitative researchers.
Hybrid quantum-classical learning pipelines combine conventional accelerators, quantum runtimes, and quantum processing units (QPUs), creating scheduling, memory, isolation, encoding, and deployment challenges that are not captured by application-level quantum machine learning surveys alone. This paper presents a systematic review of runtime and systems mechanisms for hybrid quantum-classical workloads, with medical imaging used as a translation lens rather than as an exclusive inclusion boundary. Following a PRISMA-aligned review process, we screened 364 records and synthesized 40 studies published between 2020 and 2025. Each study was coded by systems layer, application grounding, noisy-label relevance, and evaluation maturity. The coding shows that the corpus combines direct medical evidence with broader transferable systems evidence: 8 studies directly evaluated medical data, 12 were medically motivated, and 20 were generic systems studies. Across the corpus, the strongest support concerns hybrid orchestration, qubit/resource allocation, classical-quantum data movement, and container-based reproducibility, whereas evidence remains limited for realistic clinical operation, end-to-end remote-QPU workflows, multi-tenant isolation, and noisy-label retraining loops. We contribute an evidence map, a direct/indirect/interpretive evidence distinction, and cross-layer design guidelines for future hybrid quantum-classical imaging pipelines in regulated settings.
Background/Objectives: Dynamic navigation and robot-assisted implant workflows depend on robust intraoral perception. Marker-based tracking introduces workflow complexity and is sensitive to occlusions, motivating markerless alternatives. This study evaluates whether a single-stage YOLO instance segmentation model (YOLO-seg) can provide a practical markerless perception layer for dental navigation, combining accurate per-tooth delineation with low, predictable inference latency. Methods: YOLO-seg was trained end to end on an intraoral RGB corpus of 400 training, 20 validation, and 100 testing images, combining a public source and a partner-hospital in-house set. A two-stage YOLO + SAM baseline was implemented for comparison. Segmentation quality was evaluated on a 50-image held-out clinical test set at three complementary levels (per-instance matching, per-class union, and global union), with paired Wilcoxon signed-rank tests, Cliff's delta effect sizes, and 95% bootstrap confidence intervals. Runtime was assessed under matched inference-only and end-to-end conditions on N = 100 frames at a 640 × 640 resolution on an NVIDIA RTX A2000 GPU. Results: YOLO-seg significantly outperformed YOLO + SAM across all primary metrics, with very large effect sizes (Cliff's delta: 0.76-0.94; Wilcoxon p < 10-8 on every metric except precision at IoU ≥ 0.5). YOLO-seg reached AP50 = 0.716 and recall = 0.973 versus 0.383 and 0.398 for YOLO + SAM. Under matched inference-only timing, YOLO-seg ran at 27.08 ms per frame (36.9 FPS) versus 1302.78 ms (0.77 FPS), an approximately 48-fold latency gap intrinsic to the two-stage forward pass. Conclusions: YOLO-seg shows strong potential as a 2D perception module for dental navigation, balancing per-instance segmentation fidelity with real-time feasibility under the tested conditions. These results support its use as a 2D perception front-end for future integration with stereo-based 3D reconstruction and robot-assisted navigation; 3D registration accuracy, implant-placement error, and robotic execution remain outside the scope of the present study.
A sensitive and selective LC-MS/MS bioanalytical method was developed, optimized, and validated for the simultaneous quantification of bupropion and hydroxybupropion in human plasma using ultrasound-assisted dispersive liquid-liquid microextraction (UA-DLLME) as the sample preparation technique. Four critical UA-DLLME parameters - extraction solvent volume, disperser solvent volume, sample pH, and ultrasonication time - were simultaneously optimized using Central Composite Design and response surface methodology, yielding extraction recoveries of 93.8 ± 1.2% and 95.6 ± 0.9% for bupropion and hydroxybupropion, respectively. Chromatographic separation of both analytes and the internal standard carbamazepine was achieved on an Agilent Poroshell 120 EC-C18 column (50 × 4.6 mm, 2.7 μm) using an isocratic mobile phase of 10 mM ammonium formate (pH 4.0) and acetonitrile (35:65, v/v) at 0.4 mL/min, with a total runtime of 5 min. Detection was performed by triple quadrupole mass spectrometry in positive electrospray ionization mode using multiple reaction monitoring transitions of m/z 240.1 → 184.1 for bupropion and m/z 256.1 → 238.1 for hydroxybupropion. Comprehensive validation following ICH M10 guideline recommendations demonstrated linearity across 2-1000 ng/mL, LLOQ of 2 ng/mL, intraday and interday precision within 14.13% CV, accuracy within ±8.92% RE, and stability under all evaluated conditions for both analytes. The method was successfully applied to a single-dose pharmacokinetic study in five healthy volunteers following bupropion SR 150 mg administration. The assay measures total (non-stereoselective) hydroxybupropion, and the hydroxybupropion/bupropion AUC metabolic ratios of 6.55-11.68 serve as surrogate indices of CYP2B6 hydroxylation activity, classifying all subjects as CYP2B6 extensive metabolizers and demonstrating the method's utility for CYP2B6 phenotyping and precision pharmacotherapy applications.
The rising adoption of cryptocurrencies has been paralleled by the emergence of cryptojacking malware, malicious software that covertly hijacks computing resources for unauthorized cryptocurrency mining. This threat not only degrades system performance and incurs financial losses but also increasingly evades traditional intrusion detection systems (IDSs), which often suffer from limited accuracy, high false positive rates, and poor adaptability in real-time or resource-constrained environments. To overcome these limitations, this study proposes CryptoIDS-ViT, an advanced IDS framework that employs vision transformer architectures, namely ViT, MaxViT, and SwinViT, for robust host-based cryptojacking detection. The significance of this work lies in both the growing need for accurate cryptojacking mitigation and the introduction of a novel end-to-end detection framework that adapts transformer-based models to a security-critical domain. The novelty of the approach stems from its image-based malware representation pipeline, where executable binaries are systematically transformed into both color and grayscale images, and from the custom adaptation and comparative benchmarking of multiple pretrained transformer models for detecting cryptojacking attacks, an underexplored and emerging cybersecurity challenge. A two-stage process involving deep feature extraction and classification enables accurate detection with high generalizability. Extensive evaluation on a balanced dataset of 40,000 samples (20,000 cryptojacking, 20,000 benign) demonstrates superior performance: the SwinViT model achieves 99.35% accuracy on color images and 99.08% on grayscale, with a precision of 99.41%, recall of 99.27%, and F1-score of 99.34%. Compared to state-of-the-art CNN-based IDSs, CryptoIDS-ViT delivers a 3-4% improvement in detection metrics while preserving real-time inference efficiency. Additional contributions include: (i) a custom binary-to-image encoding strategy tailored to malware feature extraction, (ii) ViT attention heatmap visualizations that enhance interpretability and model transparency, (iii) an ablation study evaluating input modality effects (color vs. grayscale), and (iv) a runtime and memory complexity analysis supporting deployment feasibility in edge or enterprise systems. These results validate the proposed framework's capability to generalize across input types and deployment contexts, establishing ViT-based IDSs as a scalable, interpretable, and practical solution for detecting sophisticated cryptojacking malware in modern cybersecurity infrastructures. Unlike existing studies that apply ViTs to general malware classification, this work delivers a focused, thoroughly evaluated, and high-impact contribution to the detection of host-based cryptojacking threats.
CRISPR spacer-protospacer matching is widely used to infer host-virus interactions in microbial and viromics studies, but the choice of sequence search or alignment tool and its reporting behavior is often under-evaluated for this specific task. Using synthetic, semi-synthetic, and real datasets, we benchmarked commonly used tools and observed substantial differences in recall, runtime, and resource usage across distance metrics and thresholds. Our analyses support practical defaults for large-scale spacer-target matching and clarify trade-offs between exhaustive and heuristic approaches. Source code and benchmark workflows are available at https://github.com/UriNeri/spacer_matching_bench. Data and run artifacts are archived on Zenodo (https://doi.org/10.5281/zenodo.15171878).
Public leaderboards such as the Therapeutics Data Commons (TDC) ADMET benchmark are widely treated as a ranking of state-of-the-art models. However, a high leaderboard position is only meaningful if the corresponding model can actually be reproduced and deployed by an independent researcher. In this work, we audit whether the top-ranked TDC ADMET models meet that bar. We assessed the top-ranked models of all 22 TDC ADMET leaderboards from the perspective of an end user with access only to the publicly released artifacts of each model─its publication, code repository, and installation instructions. For every end point, the top three models were screened with a unified protocol including an execution environment reproducibility check, a data-leakage assessment, verification of the hyperparameter-optimization procedure, and a reevaluation against the current leaderboard. Only three models (CaliciBoost, MapLight, and MapLight + GNN) passed all stages and reproduced their reported performance. The remaining models failed because of unavailable code, nonreproducible environments, runtime incompatibilities, or methodological flaws. We traced direct or indirect data leakage in the MiniMol, GradientBoost, and XGBoost models, and used deliberately overfitted variants of our own Mol2Vec-based models to show that tuning on the public test set─whether accidental or intentional─can substantially inflate both metrics and leaderboard rank. These results indicate that current TDC leaderboard positions cannot be read as a direct measure of model quality and practical applicability and emphasize the urgent need for better public ADMET benchmarks based on the hidden test sets, strict data set versioning and model submission with standardized inference environments.
Secure dissemination of high-resolution satellite imagery remains challenging because many image-tailored ciphers either (i) emphasize permutation-heavy designs without sufficiently strong, plaintext-adaptive nonlinearity, or (ii) provide strong security metrics but fall short on scalable, near-real-time performance and robustness assessment under realistic channel impairments. To address these gaps, this work proposes a three-stage chaos-chess hybrid encryption pipeline for color satellite images that couples fractional-order hyperchaotic key generation with lightweight algebraic mixing, dynamic substitution, and structured bit-level diffusion. First, multiple images are optionally augmented and each RGB channel is partitioned into [Formula: see text] pixel matrices that are mixed via invertible matrices derived from a 6D fractional-order hyperchaotic Vaidyanathan system, providing efficient confusion suitable for parallelization. Second, plaintext-sensitive S-boxes are constructed online from a 4D fractional-order hyperchaotic system and applied per channel to enhance nonlinearity and satisfy stringent criteria (NL [Formula: see text], SAC [Formula: see text], low LAP and DAP). Third, the resulting bit-streams are diffused by traversing [Formula: see text] blocks using Knight's Tour paths and XORing with 4D hyperchaotic key-streams to amplify avalanche propagation. Experiments on satellite and natural images demonstrate high ciphertext randomness (entropy [Formula: see text]), strong differential resistance (NPCR [Formula: see text], UACI [Formula: see text]), near-zero adjacent-pixel correlation (PCC [Formula: see text]), and a large key space ([Formula: see text]), while measured runtimes indicate suitability for real-time or near-real-time operation. Noise-like ciphertexts and lossless recovery are verified via visual, histogram, and DFT analyses, and robustness under occlusion and noise attacks (salt-and-pepper, Gaussian) is evidenced. The resulting modular design provides a scalable pathway for protecting remote sensing data and supports future integration with ROI-aware processing and hardware acceleration.
Salt-and-pepper noise is a particularly challenging image degradation model because it corrupts pixels sparsely but severely, often disrupting edges, textures, and fine structural details. This paper presents the exponential tension field (ETF) as a geometrically grounded method for removing impulse noise. ETF is formulated from an exponential gradient-based functional and implemented as a nonlinear, gradient-dependent diffusion operator that promotes smoothing in corrupted homogeneous regions while limiting diffusion across significant intensity transitions. The proposed method is discretized at the pixel level using a conservative finite-volume scheme and evaluated on standard grayscale benchmark images degraded by salt-and-pepper noise at various levels. We assess performance using PSNR, SSIM, precision, recall, F-measure, VIF, FOM, and MAE, along with runtime analysis and statistical testing. ETF is compared to representative classical, variational, and transform-based denoisers. Benchmark results show that ETF excels at restoring images, particularly under high-noise conditions, while preserving edges and fine image structures. The method does not require training, is computationally efficient, and has a clear variational interpretation. The findings suggest that ETF is a viable and comprehensible geometry-based method for denoising salt-and-pepper images in the examined context, prompting further investigation into similar nonlinear diffusion operators for impulse-noise image restoration.
Large audio archives contain rich and diverse sonic material, yet they are seldom usable as controllable media in interactive contexts such as installations, live performance and adaptive sound environments. This paper presents a framework for interactive latent audio synthesis and technically continuous sound-space traversal and synthesis within a structured latent manifold rather than unconstrained audio generation. The framework first uses pretrained audio encoders, including AudioMAE, CLAP and related models, to organize a curated 120,000-clip AudioSet subset into a structured audio embedding space. A variational autoencoder then learns a smooth latent representation, which is further refined by a latent diffusion model to improve latent validity and traversal continuity. The refined latent codes are rendered into controllable waveforms through a DDSP-based synthesis stage, while Ambisonic spatialization provides Ambisonic spatial rendering coupled to traversal parameters. Gesture is used only at inference time as a control layer for traversal and spatial modulation, rather than as a training condition for the generator. The framework is evaluated against VAE-only and diffusion-only baselines using latent-structure analysis, interpolation behavior, synthesis quality and runtime performance. Results show that the proposed hybrid model achieves a CLAP similarity of 0.82, a mean F0 error of 245.3 Hz, a spectral convergence of 0.132 and an interactive latency of approximately 35 ms. These findings provide technical and proxy-based evidence for latent continuity, synthesis stability and real-time traversal feasibility. These findings provide technical and proxy-based evidence for latent continuity, synthesis stability and real-time traversal feasibility, while the human-centered pilot evaluation provides initial user-level evidence for perceived traversal smoothness, controllability, responsiveness and creative usefulness. Because the pilot evaluation is small-scale, these user-facing findings should be interpreted as preliminary rather than as statistically generalizable validation.
Histophilosis is an important cause of morbidity and mortality as well as antimicrobial use in feedlot cattle across North America. Detection of Histophilus somni by culture is challenging, and there is no standardized tool for distinguishing isolates that carry virulence factors most likely to contribute to disease. The DR2 repeat of H. somni-associated virulence factor 'immunoglobulin-binding protein A' (ibpA DR2) harbors a Fic domain that mediates host cell cytotoxicity and is essential for histophilosis. For rapid detection of ibpA DR2 in extracted DNA, we developed a real-time recombinase polymerase amplification (RPA) assay with a runtime of 24 min at 39 °C. DNA from H. somni-RPA-positive respiratory swabs (n = 73) was screened for ibpA DR2 using the novel RPA assay and long-read metagenomic sequencing, as well as nanopore whole-genome sequencing (WGS) of H. somni isolated from the same samples. IbpA DR2 was identified in 71% and 70% of tested samples using RPA and WGS, respectively, and in ≤41% of samples using metagenomic sequencing. The likelihood of detection by RPA did not differ (OR 1.1, 95% CI (0.42, 2.9), P > 0.99) from WGS; however, agreement between these assays was only fair (κ = 0.31). Conversely, RPA (OR 3.4, 95% CI (1.6, 8.2)) and WGS (OR 8.0, 95% CI (2.4, 42)) were more likely (P < 0.001) to detect ibpA DR2 than metagenomic sequencing, likely reflecting limited coverage of H. somni by metagenomics. This study demonstrated that RPA and long-read WGS detected ibpA DR2 with similar frequencies in extracted DNA and H. somni isolates, respectively. Further testing of non-target isolates confirmed the analytical specificity of ibpA DR2 to H. somni. Further investigation of the diagnostic validity for RPA-based ibpA DR2 detection is required in a larger cohort of field samples, as a rapid screening tool for H. somni most likely to contribute to disease.
Foundation models (FMs) promise to standardise predictive modeling across domains, yet their clinical value for tabular data remains unproven. To test this, we performed a large, fully reproducible benchmark of TabPFN, a leading FM for tabular prediction, against twelve established machine learning (ML) methods across twelve binary clinical tasks. Cohorts spanned 788 - 139,528 patients across diverse outcomes, including survival, metastasis, and disease status. Using standardized preprocessing, bootstrapping, and multiple performance metrics, TabPFN was generally competitive but did not consistently outperform strong ML baselines. It exceeded the best ML model in only 16.7% of tasks, with most area under the receiver operating characteristic (AUROC) differences within ± 0.02. TabPFN also incurred higher computational cost, with median runtimes 5.5× longer and practical reliance on GPU acceleration. These findings indicate that, for routine binary clinical prediction, TabPFN offers limited performance gains relative to optimized ML methods, while introducing significant efficiency trade-offs.
The continually increasing volume of sequence data results in a growing demand for fast implementations of core algorithms. Computation of pairwise alignments based on dynamic programming is an important part in many bioinformatics pipelines and a major contributor to overall runtime due to the associated quadratic time complexity. This motivates the need for a library of efficient implementations on modern GPUs for a variety of alignment algorithms for different types of sequence data including DNA, RNA, and proteins. Accelign is a library of accelerated pairwise sequence alignment algorithms for CUDA-enabled GPUs. Its parallelization strategy is based on a common wavefront design that can be adapted to support a variety of dynamic programming algorithms: local, global, and semi-global alignment of genomic and protein sequences with a variety of commonly used scoring schemes supporting one-to-one, one-to-many or all-to-all pairwise sequence alignments. This leads to a peak performance between 16.1 TCUPS and 9.1 TCUPS for computing optimal global alignment scores with linear gaps and affine gap penalties on a single RTX PRO 6000 Blackwell GPU, respectively. In addition, our library demonstrates significant speedups in several real-world case studies over prior CPU-based (SeqAn, Parasail, BSalign, EdLib, KSW2, WFA2, A*PA2) and GPU-based libraries (ADEPT, GASAL2), and can even outperform highly customized algorithms (WFA-GPU, CUDASW++4.0). Furthermore, the performance of our approach scales linearly with the number of employed GPUs, which makes it feasible to exploit multi-GPU nodes for increased processing speeds. Accelign provides significant speedups for commonly used pairwise alignment algorithms compared to prior implementations. It is freely available at https://github.com/fkallen/Accelign .