IEEE Transactions on Computational Biology and Bioinformatics

Learning from imbalanced data remains a major challenge for graph neural networks (GNNs), as minority nodes are not only rare but also structurally marginalized within the graph. We address this issue with CLARA, a hierarchical learning framework that decomposes node classification into two stages: a coarse subgraph-level classifier that selects regions likely to contain minority instances, followed by a fine-grained node-level predictor within these regions. This design improves sensitivity while maintaining scalability, filtering out irrelevant areas and focusing learning on topologically meaningful neighborhoods. Experiments on benchmark graph datasets demonstrate substantial gains over established imbalance-handling methods, with CLARA reaching an F1-score of 88.3\%. The same strategy achieves significant improvements in protein-ligand binding site prediction, underscoring its broad and consistent effectiveness across both biological and general graph learning tasks.

Graph embedding transforms complex networks into low-dimensional representations for analysis and downstream machine learning. Existing methods often rely on eigenvalue ordering, random-walk sampling, or neural optimization. We introduce MOSAIC, Multistep Operator Spectral Accumulation with Isolated Components, an unsupervised framework that constructs an accumulated diffusion operator from successive powers of a graph transition matrix. MOSAIC uses Krylov subspace analysis to identify dominant spectral components according to their convergence behavior. This approach emphasizes persistent and spectrally separated diffusion modes rather than selecting embedding coordinates solely by eigenvalue magnitude. We evaluated MOSAIC using a biomedical graph projection derived from the Data Distillery Knowledge Graph that contained 5,171 human genes and 10,147 phenotypes. Across embedding dimensions from 16 to 128, MOSAIC achieved node-type classification AUROCs from 0.9863 to 0.9926 and exceeded FastRP, Node2Vec, HashGNN, and GraphSAGE. For gene-phenotype link prediction, validation and test associations were withheld before embedding generation. Under progressive removal of all edges or specific edge classes, MOSAIC retained more node-label information than FastRP, including after extreme graph sparsification. In a focused DICER1 analysis, 106 of 137 documented phenotype associations received scores above 0.90, demonstrating concordance with known pleiotropic biology and prioritizing additional relationships for independent evaluation. These results support MOSAIC as a fully unsupervised framework for learning informative graph representations. Although demonstrated using a biomedical graph projection, MOSAIC is domain independent and can be applied to other complex networks

Protein-protein interactions shape cellular organization and disease processes, but their prediction remains difficult because known interaction networks are incomplete, noisy and highly uneven across proteins. Hyperbolic graph learning is well suited to the hierarchical structure of such networks, yet most methods emphasize learning hyperbolic embeddings and pay less attention to how these embeddings are propagated and decoded for type prediction. Here we introduce the Lorentz Tangent Diffusion Adapter (LTDA), a post-embedding module for PPI type prediction, that is, for predicting which of several interaction types a candidate protein pair undergoes. LTDA learns relation-specific protein representations on the Lorentz manifold, projects them to tangent space, applies lightweight graph diffusion, and decodes the diffused representations directly in tangent space. This design uses the tangent space as a prediction-oriented interface between hierarchical representation learning and interaction classification. On SHS27K and SHS148K under BFS and DFS splits, LTDA consistently improves over representative PPI methods and a reproduced HI-PPI baseline, with average gains of 1.41 percentage points in Micro-F1, 3.21 in AUPR, 1.31 in AUC and 1.25 in accuracy. Ablation, degree-bin and biological-evidence analyses indicate that the gains are not explained by simple smoothing or hub correction. LTDA therefore shows that post-embedding adaptation can make Lorentz representations more effective for PPI type prediction.

Drug repurposing and target discovery offer critical strategies for advancing therapeutic development by uncovering the potential biological pathways and novel associations among drugs, genes, and diseases. However, experimental discovery remains expensive and time-consuming, which limits the scalability of large-scale studies. In addition, existing computational approaches often struggle to effectively integrate heterogeneous biomedical data, capture the complex higher-order topological signatures of biological interactomes, and generalize to unseen entities. Here, we present HANAMI (Heterogeneous grAph coNtrastive leArning for drug-gene-disease Motif predIction), a multi-view deep graph learning framework designed to model complex interactions among drugs, genes, and diseases. HANAMI integrates diverse heterogeneous biomedical knowledge, including chemical structures, genomic sequences, and clinical phenotypes, and leverages relation-aware topology encoding, structure-aware aggregation, and contrastive learning to enable accurate motif prediction with biological context from the network. Systematic evaluation on benchmark datasets shows that HANAMI achieves up to 6% improvements over existing state-of-the-art methods in predicting drug-gene-disease motifs. The framework further demonstrates strong inductive generalization, maintaining an [~]18% performance advantage in zero-shot settings involving previously unseen entities. Beyond predictive performance, HANAMI effectively prioritizes drug-disease relationships investigated in Phase II or III trials while identifying candidate genes that suggest plausible mechanistic links. Together, HANAMI provides a computational framework for interpreting complex biomedical interactions, offering a scalable foundation to accelerate drug repurposing and therapeutic innovation.

Genotype-by-environment interactions can complicate plant breeding because genetic effects may vary across weather regimes, locations, and years. Although genomic prediction models can use dense marker data to predict complex crop traits, they often provide limited information about how marker relevance changes across environmental conditions. This study introduces a graph-based framework for estimating environment-associated marker rankings from multi-environment crop trial data. We introduce an attention-based graph neural network that jointly represents SNP markers and weather-based environment vectors as interacting nodes and is designed to learn environment-associated marker rankings from supervised yield data. The model is trained using yield prediction as the supervised objective, and the learned marker-level attention signals are aggregated to estimate marker rankings overall and within groups of similar environments. We evaluate the framework on two genotype-by-environment datasets. In both datasets, we examine whether learned marker rankings are stable across independent training runs and determine the overlap between high-ranked markers and external GWAS references. In a maize dataset with data from 212 different site-years, we rank markers separately within each weather-defined environment cluster and define a global marker set as the markers shared across all clusters. This provides a structured way to examine how marker rankings vary across weather-defined groups and which markers remain important across environments. We further encode each weather variable with a separate LSTM combined by attention, which makes the contribution of individual weather variables readable, and compare the resulting variable ranking against an independent SHAP-based attribution. Genes located near the top-ranked markers are also examined for functional enrichment. Overall, the results suggest that graph-based, environment-conditioned marker attribution may complement genomic prediction as an exploratory tool for studying context-dependent genomic signals in multi-environment trials.

Accurate molecular property prediction requires effective molecular representations that can describe a molecule from multiple complementary perspectives. Existing deep learning approaches typically use SMILES strings, two-dimensional molecular graphs, or three-dimensional conformations as inputs. These representations capture different aspects of molecular information: SMILES encodes a sequential description, molecular graphs describe atom-bond connectivity, and 3D conformations provide spatial information from atomic coordinates. However, these modalities are often learned separately or merged with low effective fusion operations, which may not sufficiently capture interactions among sequential, topological, and geometric features. In this work, we propose Tri-Modality Cross-Attention (TMCA), a multimodal framework that integrates 1D SMILES, 2D molecular graphs, and 3D molecular conformations for molecular property prediction. TMCA uses pretrained encoders for the 1D and 3D branches, with a SMILES-based Transformer for the 1D branch and a recent conformation-aware pretrained model for the 3D branch. At the same time, a trainable 2D graph encoder is designed to support modality fusion. All the parameters from the pretrained models are frozen to make overall training more efficient. We evaluate TMCA on four datasets (i.e., BBBP, BACE, ClinTox, and HIV) from the MoleculeNet database for classification tasks and compare its performance with state-of-the-art competing approaches. Experimental results show that TMCA achieves the best overall performance, demonstrating the value of multi-modality integration for drug property prediction and the effectiveness of our cross-attention-based fusion strategy. The source code of TMCA is available at: https://github.com/Ay-Zhao/TMCA

In the recent growing interest of AI research toward biology, genetic association studies of AI- derived phenotypes from high-content modalities such as images emerges as a powerful means for biological discovery. However, such AI-phenotyping methods still lacks a good optimization target and an efficient evaluation framework. The number of discovered loci was the major criterion for evaluating the quality of AI-derived phenotypes. However, the time and computational resources required for running the GWAS and subsequent loci-clumping are substantial, limiting the rapid development and iteration of deep learning representation algorithms. Here we present a 1000x faster and lightweight framework, sum-h2, than traditional GWAS framework for evaluating genetic discovery through total heritability. We revisit tr(P^(-1) G), previously proposed in the context of evolutionary studies, as a measure of multi-phenotype heritability. We showed that sum-h2, the sum of heritability over phenotypic Principal Components (PCs), is equivalent to the linear transformation-invariant tr(P^(-1) G), through both theoretical proof and simulation studies. Moreover, sum-h2 can be estimated rapidly with minimal information loss over a relatedness-enriched sample, while preserving the relative ranking of endophenotypes by GWAS loci counts. Based on selected UKB data and sum-h2, we set up a Genetic Discovery Arena, enabling rapid and fair comparisons for the development and optimization of deep learning-derived phenotyping methods.

Background and Objective: Prescribing multiple drugs simultaneously, known as polypharmacy, is increasingly common in managing chronic diseases such as diabetes and cardiovascular conditions. Studies show that roughly 40% of elderly patients are on five or more medications, putting them at significant risk of adverse drug-drug interactions (DDIs) [1]. While computational methods based on Graph Neural Networks (GNNs) have shown strong potential for predicting these interactions from a drugs molecular structure, most existing models are limited in two important ways: they rely on attention mechanisms that cannot adapt to the unique context of each atom in a molecule, and they only predict whether an interaction exists rather than how dangerous it might be. Methods: We propose HMVA-DDI, a two-stage deep learning framework that addresses both shortcomings. In the first stage, our model predicts whether two drugs will interact. If it detects an interaction, the second stage classifies it into one of three clinical severity levels: minor, moderate, or major. To better capture molecular structure, we replace the standard attention mechanism with a more powerful version called Graph Attention Network v2 (GATv2), which considers each atoms specific context when learning its representation. We also introduce a technique to prevent information loss in deeper network layers, and we tailor the input features separately for each prediction task. Results: Evaluated on two standard benchmarks, DrugBank 5.0.3 [2] and DDInter [3], HMVA-DDI achieves an AUROC of 0.9819 and AUPRC of 0.9759 on binary interaction detection, a marked improvement over the baseline. For severity classification, our best model variant (qcSvr4) reaches an overall AUROC of 0.9444, with a particularly large improvement in detecting low-severity (Minor) interactions, a historically difficult class. Conclusion: HMVA-DDI bridges the gap between computational drug research and clinical usefulness by not only predicting interactions accurately but also indicating interaction severity. This two-level system can help healthcare professionals prioritize warnings, reducing the alert fatigue problem where doctors are overwhelmed by too many non-critical notifications.

Conventional short read DNA alignment algorithms, such as widely used BWA-MEM2, devote a substantial fraction of their execution time to seed generation. We present RosaSeed, a short read alignment algorithm, based on a configurable seeding framework, that significantly accelerates alignment while maintaining accuracy comparable to that of BWA-MEM2. RosaSeed replaces the BWA-MEM2 seeding kernel and directly supplies candidate seeds to the existing BWA-MEM2 chaining and alignment extension pipeline, producing standard SAM output suitable for downstream variant analysis. RosaSeed was evaluated on both simulated and real short read datasets, and its seeding time, fastq-to-SAM time, alignment accuracy, and memory footprint were compared against BWA-MEM2, Minimap2, Bowtie2, and the Enumerated Radix Tree (ERT) algorithm. We also compared against ERT2, a proposed simplified variant of ERT that sacrifices a small amount of alignment accuracy to significantly increase the alignment throughput. With single-threaded software execution, our recommended RosaSeed configuration achieves 13.06x and 3.85x speedups in seed processing and fastq-to-SAM time, respectively, over BWA-MEM2. Corresponding seed processing (and fastq-to-SAM) speedups are 3.96x (2.48x) over ERT, 1.2x (1.44x) over ERT2, 4.82x (3.31x) over Minimap2, and 33.24x (10.83x) over Bowtie2. RosaSeed requires only 49.61 GB of peak memory, approximately 25% less than ERT and ERT2, while maintaining alignment accuracy comparable to that of BWA-MEM2. Following the recent release of minibwa, a prefetch-optimized FM index aligner, we additionally benchmarked RosaSeed against it on an upgraded 24-core, 48-thread AMD Zen 3 workstation. Our speed-optimized miniRosaSeed configuration achieves a 2.12x fastq-to-SAM speedup over minibwa for single-threaded execution while producing 2.58% higher standard accuracy. Against Strobealign, miniRosaSeed reported a 2.18x fastq-to-SAM speedup and 3.1% higher standard accuracy. For multi-threaded execution, miniRosaSeed was both faster and more accurate than minibwa for up to 26 software threads on the Zen 3 workstation. RosaSeed provides multiple runtime-memory configurations, spanning memory-efficient and high-performance strategies, enabling users to select operating points appropriate for diverse computational environments while maintaining alignment quality comparable to that of BWA-MEM2.

Local ancestry inference (LAI) reveals population compositions of an individuals chromo-somes, which empowers population genetics and genetic association studies of diverse populations. The recent availability of large genomic datasets from biobanks offers unprecedented opportunities for LAI. However, such datasets are typically only accessible from cloud-based platforms, and the lack of efficient and cost-effective tooling makes LAI prohibitive for resource-constrained researchers. Here, we develop Recomb-Mix Cloud, a cloud-ready pipeline for LAI based on Recomb-Mix, a recently introduced efficient LAI method. Recomb-Mix focuses on individual variants while maintaining a lower computational burden than other LAI methods focusing on genomic windows. Recomb-Mix Cloud offers a cloud-friendly complete pipeline from panel sample selection, marker alignment, compact reference panel construction, Recomb-Mix threading, and generating LAI calls in common data formats. Recomb-Mix Cloud also offers a set of pre-computed compact population reference panels constructed based on data from the Human Genome Diversity Project and the 1000 Genomes Project. Our results showed Recomb-Mix is about one and two orders of magnitude cheaper and faster than FLARE and RFMix, two popular LAI methods, while offering largely consistent LAI calls on the All of Us Researcher Workbench. The estimated overall compute cost per All of Us sample for whole-genome LAI analysis using Recomb-Mix Cloud with a pre-computed reference panel is just $0.01, compared with $0.43 for FLARE and $4.58 for RFMix. Therefore, Recomb-Mix Cloud is well-positioned to democratize local ancestry-based genetics research on modern biobank-sized data. The Recomb-Mix Cloud pipeline and the compact reference panels are available at https://github.com/ucfcbb/Recomb-Mix.

Biomarker discovery from high-dimensional RNA sequencing data remains challenging. Conventional methods such as differential expression analysis, pairwise protein-protein interaction networks, and weighted gene co-expression network analysis suffer from false-positive interactions, transitivity-driven noise, and arbitrary thresholds for feature selection. Addressing these limitations, this study proposed a novel framework, Hypergraph Regularised Fuzzy Rough Gene Selection (HR-FRGS), for knowledge-driven biomarker discovery. This pipeline integrates information-rich protein clustering with entropy filtering and PPI-pruned co-expression interactions to form a dual-layer hypergraph. A hypergraph-based scoring method that combines Random Walk with Restart diffusion and Hypergraph Betweenness Centrality is used to identify genes that are central in the biological network. Furthermore an autonomous fuzzy rough set selection eliminates arbitrary threshold dependency for gene selection. HR-FRGS was applied to four TCGA cancer cohorts: Lung Adenocarcinoma, Head and Neck Cancer, Kidney Clear Cell Carcinoma, and Colon Cancer, which reduced over 60,000 transcripts to compact biomarker panels. Validation using classical Machine Learning algorithms with cross-validation demonstrated that biomarker panels matched or exceeded the classification performance of the full transcriptome (AUC > 0.99, MCC > 0.94) and confirmed generalizability (MCC $\geq$ 0.893). Benchmarking against baseline methods and an ablation study showed the contribution of major parts of the algorithm. Functional enrichment analysis captures established pan-cancer hallmarks and cohort-specific oncogenic mechanisms, confirming the biological relevance of the identified genes.

Integrative analysis of the properties of multiple networks pertaining to a biological system is a key problem in systems biology. But computing the properties of thousands of large networks poses a challenge. Most techniques that partially address this challenge, such as parallelism and optimized algorithms, treat the analysis of each network as an independent task. However, the networks arising from the same biological system are often similar to each other, and can we harness this similarity to achieve speedup? Towards this end, we propose a phylogeny-like data structure, nc-tree, to compactly represent an ensemble of related biological networks; and a versatile framework, ProcessNets, that runs incremental algorithms over the nc-tree to speedup property computations on the input networks. Our asymptotic analysis of space and time complexity, coupled with empirical evaluations of diverse ensembles of simulated and real-world GTEx gene coexpression networks, demonstrate that ProcessNets leads to significant space and time advantages when computing different network science measures on ensembles of similar networks. For instance, we achieve a compression factor of 2.1x and 8.1x in storing two representative ensembles of GTEx Muscle Skeletal networks with 1000 nodes/genes, and a time speedup of 2.86x to 3.97x over the fastest baseline for computing degree centrality in ensembles of networks with at least 14,500 nodes derived from subsampled GTEx datasets of ten tissues. These results are promising and encourage application of ProcessNets to analyze biological network ensembles derived from rapidly accumulating consortium/biobank-based datasets.

Tabular foundation models are used as off-the-shelf predictors for heterogeneous tabular tasks, but it remains unclear how they will perform on real genotype-to-phenotype tabular datasets, which carry unique challenges. One such challenge is ancestry-dependent non-stationarity in allele effect sizes (the effects of genetic variation on disease risk): the predictive contribution of a given genetic variant may change significantly across the ancestry spectrum. This is especially consequential for patients from ancestry groups that are underrepresented in existing datasets, because models that are unaware of this non-stationarity are more likely to perform poorly for them. The battery of synthetic tasks used to pretrain tabular foundation models does not capture this structure, and we show that it leads to systematic performance degradation. Using controlled hierarchical Gaussian-process stress tests, we demonstrate that both off-the-shelf TabICL and TabPFN are robust when ancestry dispersion in the data is low, but as dispersion grows, the latent ancestry-dependent non-stationarity of effect sizes becomes a first-order problem, and the predictive performance of both models deteriorates. We confirm the same pattern on All of Us (AoU) (a large biobank containing whole-genome sequencing and electronic health record (EHR) data spanning the ancestry spectrum) across an extensive panel of cancer phenotypes evaluated with the two leading tabular foundation models: TabICL and TabPFN. Holding the in-context training-set size, phenotype, and feature set fixed, ancestry-specific in-context tables, which are less dispersed in ancestry space, consistently outperform size-matched meta-ancestry tables, isolating ancestry dispersion as the driver of degradation. To address the failure without changing the base architecture, we construct two synthetic task families that explicitly encode ancestry-dependent effect drift and instruction-tune an off-the-shelf TabICL model on tasks sampled across both families. On held-out AoU evaluations covering cancer and respiratory disease phenotypes, the tuned model yields more stable performance across ancestry-distance bins and is especially strong in the bins farthest from the center of the in-context exemplars in ancestry space, that is, on subjects whose ancestry is most underrepresented among the provided exemplars. The paper contributes an evaluation protocol, a failure analysis, and two synthetic task families targeted at ancestry-dependent non-stationarity in precision-medicine tabular modeling.

In network biology, analyzing pathways is critical to study mechanisms or sequences of biological events. The pathway co-membership approach aims to link proteins belonging to the same pathway inside a network that is a clique. Although this method is great for projecting pathway knowledge onto proteins, it remains unsuitable for some downstream applications relying on graph distances like random walks with restart. To address this limitation, we introduce a weighted pathway co-membership method that incorporates biological distances between proteins. We achieve this by developing an abstraction framework for Reactome pathways, which captures both their hierarchical organization and sequential relationships. By assigning weights to these abstractions, we define a scoring system for proteins within the pathway co-membership clique. We illustrate our approach using a Reactome pathway ("Signaling by TGF-beta receptor complex", R-HSA-170834), and explain how we generate the weighted co-membership cliques of the 29 Reactome top-level pathways. Next, we construct the entire weighted pathway co membership network of the Reactome knowledge base. Validation and evaluation of our method is performed using cancer hallmark gene sets and a random walk with restart experiment, comparing our approach to state-of-the-art methods. We show that our approach outperforms state-of-the-art methods for gene prioritization on cancer hallmarks.

Tabular synthetic data generation (SDG) can facilitate research and development in the health domain, as access to health data is restricted. To accommodate for its structural complexity, such as multimodal distributions and fragmentation, transformer models provides an alternative to conventional SDG methods since they are better equipped to infer global spatial coherence due to textual context of tabular records. However, methods for optimising their adaption to tabular SDG are still largely unexplored. We contribute to the understanding of transformer-based SDG by evaluating fidelity and privacy of three prompting regimes, zero shot, one shot and few shot training, for tabular health SDG using small transformer models. We evaluated llama 3.2 (1B and 3B parameters) and Phi (1.3B and 2.7B parameters), on a real clinical dataset. Our results showed that zero shot learning was less prone to hallucinations and performed overall better for fidelity, but at the loss of privacy in comparison to the other regimes. Our results showed no differences between model sizes and families. Future studies should extend the analysis to other models and datasets, and explore strategies to reduce hallucinations.

We combine GeneNetwork's 1TB structured database with its metadata as an RDF graph and an associated literature corpus. We benchmark frontier and local LLMs querying it through four architectures: document, graph and agentic retrieval-augmented generation (RAG), and a combination of the three. We evaluate them on 40 domain specific expert-curated questions with a new metric: the satisfaction frequency psi (psi, i.e, the share of questions answered with precision and recall of at least 70% or an f1 of at least 60%). psi ranks systems the same way precision, recall and f1 -- a machine learning metric built from the harmonic mean of precision and recall -- but is easier to read and assess for humans. System architecture had a large effect: with full use of Claude Opus 4.8 the plain document RAG of our earlier GNQA system had a psi of 36% against 86% for our new combined tools. Delegating most of the work from Claude Opus 4.8 to a local model halved query costs while still reaching a psi of 59%.

We present a lossless compression algorithm for indexing a protein database while supporting fast taxonomic classification in the method Centrifuger. The algorithm is a new scheme of the previously proposed run-block compression algorithm to reduce the size of the Ferragina-Manzini (FM) index, and it scales better with alphabet size than the original. On the RefSeq prokaryotic and viral protein sequences, Centrifuger reduces the memory footprint by over a third compared to the method Kaiju that builds on a plain FM-index, while having comparable running time. Furthermore, the compressed FM-index is lossless and can locate matches of arbitrary length, which helps Centrifuger achieve greater accuracy than Kraken2, a k-mer-based taxonomic classification method. We leverage the computational efficiency of Centrifuger to create an index of size 182 GB for classifying the reads against the full nr database that contains about 250 billion amino acid characters. Using this index, Centrifuger reveals different SARS-CoV-2 infection states and viral transcriptome profiles across human cell types from single-cell RNA-seq data.

Interest in bacteriophages as alternative treatments for bacterial infections has grown amid the ongoing crisis of multidrug resistance. However, their narrow host range makes it labor-intensive to identify matching phages for bacterial isolates using conventional experimental methods. Although existing computational models based on biomolecular features have shown promise for phage-host matching, the performance of alternative model settings, including data embeddings, across datasets with different degrees of novelty remains underexplored. This study introduces PhageHost-GNN, a novelty-aware framework using routed models to predict strain-level Klebsiella-phage interactions, evaluated under three clinical scenarios. Using multiple embedding models to encode sequence, structural, and genomic-context features, the framework achieved AUPRCs reaching 0.784, 0.731, and 0.642 in host-unseen, phage-unseen, and both-unseen in silico evaluations, respectively. Benchmarks on experimentally validated datasets showed the clearest advantage over fixed embeddings when both hosts and phages were new to the trained model. In addition to direct inference and de novo training, the framework further implements a fine-tuning pipeline to adapt existing models to local custom datasets while reducing model-training time. Overall, PhageHost-GNN provides a flexible framework for model selection, evaluation, and local adaptation in practical phage-therapy screening.

Metagenomic sequencing enables broad pathogen detection without requiring prior assumptions about the organisms present, but the memory and computational demands of classifying large sequencing datasets against reference databases often lead analysis pipelines to rely on desktop or server-grade computing platforms. This limits practical use in disasterstricken and other resource-constrained environments, where portable sequencing may be available but reliable access to cloud or laboratory infrastructure is not. We present KENT, a portable edge-computing framework for metagenomic pathogen screening based on a cluster of NVIDIA Jetson Nano modules. KENT adapts CuCLARK-l for low-memory embedded execution and extends it with a custom software stack for distributed read classification, abundance estimation, split-run abundance merging, and report generation. Using an MPI-based controller/responder architecture, KENT supports local, offline analysis with sample-level parallelism across multiple nodes. We evaluate KENT using long read wastewater metagenomic samples with a custom pathogen database and the CAMI2 Marine long read dataset. In the wastewater setting, KENT achieves an average speedup of 3.1x over the CPU-based CLARK-l baseline while maintaining close agreement in classified read abundance estimates. On CAMI2 Marine, KENT achieves an average speedup of 2.68x and similarly preserves CLARK-l abundance profiles. Evaluation against the CAMI2 long read taxonomic ground truth yields a mean species-level precision of 58.55%, recall of 93.36%, F1 score of 71.74%, and L1 abundance distance of 0.731, closely matching CLARK-l. Among ground truth species represented in the reference database, KENT detects 99.86% on average. A controlled database-size experiment further shows little change in classification throughput across the tested database configurations, while large sequencing inputs remain constrained by the Jetson Nanos unified memory capacity. Overall, KENT demonstrates that GPU-accelerated metagenomic classification can be adapted to low-memory embedded hardware while preserving the biological profiling behavior of the CLARK-l baseline, supporting portable and offline pathogen screening in resource-constrained settings. AvailabilityKENT is available at https://github.com/BilkentCompGen/KENT

Living clinical guidelines revise recommendations as evidence emerges, so an AI answer can be faithful to a real guideline and still be out of date. Here we examine this failure using ASCOBench, 288 unique questions in 96 three-turn conversations grounded in versioned American Society of Clinical Oncology (ASCO) breast and prostate cancer guidelines, with oncologist-reviewed reference answers. Oncologists corrected one in four version-sensitive answers drafted by a frontier model, against fewer than one in ten factual ones. Across the guideline corpus, 19 recommendations changed between versions, seven of them reversals, yet only 1 of 14 superseded documents states that it has been replaced. With a strong model, retrieval over several guideline versions made stale answers four times more frequent than answering without retrieval. SentryLine, a verification-first system, lowered incorrect answers to 4.2-5.2% from 9.7-22.9% for baselines across three models. Recognizing change before a guideline declares it remains an open problem.