Large-scale Datasets Research Articles

Gene content of seawater microbes is a strong predictor of water chemistry across the Great Barrier Reef

Abstract Background Seawater microbes (bacteria and archaea) play essential roles in coral reefs by facilitating nutrient cycling, energy transfer, and overall reef ecosystem functioning. However, environmental disturbances such as degraded water quality and marine heatwaves, can impact these vital functions as seawater microbial communities experience notable shifts in composition and function when exposed to stressors. This sensitivity highlights the potential of seawater microbes to be used as indicators of reef health. Microbial indicator analysis has centered around measuring the taxonomic composition of seawater microbial communities, but this can obscure heterogeneity of gene content between taxonomically similar microbes, and thus, microbial functional genes have been hypothesized to have more scope for predictive potential, though empirical validation for this hypothesis is still pending. Using a metagenomics study framework, we establish a functional baseline of seawater microbiomes across offshore Great Barrier Reef (GBR) sites to compare the diagnostic value between taxonomic and functional information in inferring continuous physico-chemical metrics in the surrounding reef. Results Integrating gene-centric metagenomics analyses with 17 physico-chemical variables (temperature, salinity, and particulate and dissolved nutrients) across 48 reefs revealed that associations between microbial functions and environmental parameters were twice as stable compared to taxonomy-environment associations. Distinct seasonal variations in surface water chemistry were observed, with nutrient concentrations up to threefold higher during austral summer, explained by enhanced production of particulate organic matter (POM) by photoautotrophic picocyanobacteria, primarily Synechococcus. In contrast, nutrient levels were lower in winter, and POM production was also attributed to Prochlorococcus. Additionally, heterotrophic microbes (e.g., Rhodospirillaceae, Burkholderiaceae, Flavobacteriaceae, and Rhodobacteraceae) were enriched in reefs with elevated dissolved organic carbon (DOC) and phytoplankton-derived POM, encoding functional genes related to membrane transport, sugar utilization, and energy metabolism. These microbes likely contribute to the coral reef microbial loop by capturing and recycling nutrients derived from Synechococcus and Prochlorococcus, ultimately transferring nutrients from picocyanobacterial primary producers to higher trophic levels. Conclusion This study reveals that functional information in reef-associated seawater microbes more robustly associates with physico-chemical variables than taxonomic data, highlighting the importance of incorporating microbial function in reef monitoring initiatives. Our integrative approach to mine for stable seawater microbial biomarkers can be expanded to include additional continuous metrics of reef health (e.g., benthic cover of corals and macroalgae, fish counts/biomass) and may be applicable to other large-scale reef metagenomics datasets beyond the GBR.

AT4CTIRE: Adversarial Training for Cyber Threat Intelligence Relation Extraction

Cyber Threat Intelligence (CTI) plays a crucial role in cybersecurity. However, traditional information extraction has low accuracy due to the specialization of CTI and the concealment of relations. To improve the performance of CTI relation extraction in the knowledge graph, we propose a relation extraction architecture called Adversarial Training for Cyber Threat Intelligence Relation Extraction (AT4CTIRE). Additionally, we developed a large-scale cybersecurity dataset for CTI analysis and evaluation called Cyber Threat Intelligence Analysis (CTIA). Inspired by Generative Adversarial Networks, we integrate contextual semantics to refine our study. Firstly, we use some wrong triples with incorrect relations to train the generator and produce high-quality generated triples as adversarial samples. Secondly, the discriminator used actual and generated samples as training data. Integrating the discriminator and the context-embedding module facilitates a deeper understanding of contextual CTI within threat triples. Finally, training a discriminator identified the relation between the threat entities. Experimentally, we set two CTI datasets and only one baseline that we could find to test the effect in the cybersecurity domain. We also took general knowledge graph completion tests. The results demonstrate that AT4CTIRE outperforms existing methods with improved extraction accuracy and a remarkable expedited training convergence rate.

On the Practicability of Deep Learning based Anomaly Detection for Modern Online Software Systems: A Pre-Train-and-Align Framework

Operation and maintenance are critical activities in the whole life cycle of modern online software systems, and anomaly detection is a crucial step of these activities. Recent studies mainly develop deep learning techniques to complete this task. Notably, though these techniques have achieved promising results in experimental evaluations, there are still several practicality gaps for them to be successfully applied in a real-world online system, including the scalability gap, availability gap and alignment gap. To bridge these gaps, we propose an anomaly detection framework, namely ShareAD , based on a pre-train-and-align paradigm. Specifically, we argue that pre-training a shared model for anomaly detection is an effective way to bridge the scalability gap and the availability gap. To support this argument, we systematically study the necessity and feasibility of model sharing for online system maintenance. We further propose a novel model based upon Transformer encoder layers and Base layers, which works well for anomaly detection pre-training. Then, to bridge the alignment gap, we propose ShareAD alignment to align the pre-trained model with operator preference by jointly considering the local observation context and sensitivity of each monitor entity. Extensive experiments on two real-world large-scale datasets demonstrate the effectiveness and practicality of ShareAD .

Framework Using Multicriteria Analysis for Evaluating the Risk of Musculoskeletal Disorders

This study includes musculoskeletal disorder (MSD) risk evaluation based on the IMU sensor data gathered from patient-lifting movement performed by healthcare specialists. This is a continuation of previous research focusing on a novel multicriteria statistical model integrating experimental and large-scale statistical datasets. The proposed model estimates MSD probabilities over 5, 10, and 15 years for the neck (0.537 ± 0.156), shoulder (0.449 ± 0.084), and elbows (0.277 ± 0.221). The model enables individual risk profiling, influenced by dynamic parameters that can reduce the long-term risk by up to 70.49%. The model is in its early development stages, i.e., it is the proof of concept that offers a new approach to assessing MSD risk at work using motion tracking data in combination with statistics. Further studies with larger sample sizes and validated criterion weights are needed to refine and validate this approach.

<b>Machine Learning Enabled In-Home ECG: A Review</b>

Machine learning (ML)-based in-home electrocardiogram (ECG) systems have emerged as transformative tools, advancing beyond traditional cardiology methods by offering innovative techniques for cardiac care. These systems enable sustained data collection, real-time monitoring of cardiac status, and individualized treatment plans, all while minimizing the need for frequent clinic visits. By leveraging advanced analytics and ML algorithms, in-home ECG systems analyze large-scale datasets to detect patterns and anomalies that might otherwise go unnoticed, providing early alerts and improving patient outcomes. This review examines the latest trends in ML-enhanced in-home ECG technology, emphasizing its functionality in anomaly detection, continuous monitoring, and decision-making processes. The integration of ML not only enhances diagnostic precision but also opens avenues for scalable, personalized, and remote healthcare solutions. Despite these advancements, significant challenges remain, including issues related to data privacy, algorithmic biases, and the reliability of real-world implementations. Addressing these challenges is essential for optimizing the performance and ethical use of these systems. This review also explores opportunities for future research, particularly in improving algorithm robustness and addressing biases to ensure equitable and accurate cardiac care for diverse populations. By integrating state-of-the-art ML techniques, in-home ECG systems are poised to revolutionize contemporary cardiology, reducing healthcare costs and enabling a progressive shift toward accessible, patient-centered care. This comprehensive exploration highlights the potential of ML-based in-home ECG systems to redefine cardiac monitoring and treatment, contributing to the broader transformation of modern healthcare. Received: 13 September 2024 | Revised: 11 December 2024 | Accepted: 31 December 2024 Conflicts of Interest The authors declare that they have no conflicts of interest to this work. Data Availability Statement Data sharing is not applicable to this article as no new data were created or analyzed in this study. Author Contribution Statement Aqsa Bibi: Conceptualization, Methodology, Validation, Formal analysis, Investigation, Resources, Data curation, Writing - original draft, Writing - review & editing, Visualization, Project administration. Jawwad Sami Ur Rahman: Methodology, Validation, Investigation, Resources, Writing - review & editing, Supervision, Project administration.

Hybrid Offset Position Encoding for Large-Scale Point Cloud Semantic Segmentation

In recent years, large-scale point cloud semantic segmentation has been widely applied in various fields, such as remote sensing and autonomous driving. Most existing point cloud networks use local aggregation to abstract unordered point clouds layer by layer. Among these, position embedding serves as a crucial step. However, current methods of position embedding have limitations in modeling spatial relationships, especially in deeper encoders where richer spatial positional relationships are needed. To address these issues, this paper summarizes the advantages and disadvantages of mainstream position embedding methods and proposes a novel Hybrid Offset Position Encoding (HOPE) module. This module comprises two branches that compute relative positional encoding (RPE) and offset positional encoding (OPE). RPE combines explicit encoding to enhance position features through attention, learning position bias implicitly, while OPE calculates absolute position offset encoding by considering differences with grouping embeddings. These two encodings are adaptively mixed in the final output. The experiment conducted on multiple datasets demonstrates that our module helps the deep encoders of the network capture more robust features, thereby improving model performance on various baseline models. For instance, PointNet++ and PointMetaBase enhanced with HOPE achieved mIoU gains of 2.1% and 1.3% on the large-scale indoor dataset S3DIS area-5, 2.5% and 1.1% on S3DIS 6-fold, and 1.5% and 0.6% on ScanNet, respectively. RandLA-Net with HOPE achieved a 1.4% improvement on the large-scale outdoor dataset Toronto3D, all with minimal additional computational cost. PointNet++ and PointMetaBase had approximately only a 0.1 M parameter increase. This module can serve as an alternative for position embedding, and is suitable for point-based networks requiring local aggregation.

SOM modelling and prospectivity mapping of surface geochemical data in Au and multi-metal mineral exploration: example from northern Finland

Regional till and weathered bedrock geochemical data sets, provide a basis for data analysis and modelling in the glaciated terrains. These large-scale surface geochemical data sets have great potential in mineral exploration, especially when machine learning and clustering methods are used to reduce the dimensionality of multivariate data sets. Here, self-organising maps (SOM) followed by the k-means clustering were used to create SOM of the target areas for initial modelling and prospectivity mapping of Au prospecting in central Lapland, northern Finland. Due to the nature of the till and weathered bedrock data being legacy data, levelling effort was made between the map sheets. The targeting till geochemical data set did not contain the typical indicator elements for gold. Instead, elements associated with the gold deposits, within the study area, were used. Indicator associations in this study were Ni-Co with possible Cu. Resulting elemental clusters from SOM were assigned as interesting clusters according to their distribution of elements. For the till, two potential clusters were deemed, Ni-Co-Cr and Cu-V-Co. For the weathered bedrock, three clusters were specified, Ni-Co-Cr, V-Cu, and Cu-Co. This study shows the potential of using legacy data sets for early targeting stages of mineral exploration, potentially reducing the footprint of mineral exploration.

Genome-wide association study on color-image-based convolutional neural networks

Background Convolutional neural networks have excellent modeling abilities to complex large-scale datasets and have been applied to genomics. It requires converting genotype data to image format when employing convolutional neural networks to genome-wide association studies. Existing studies converting the data into grayscale images have shown promising. However, the grayscale image may cause the loss of information of the genotype data. Methods In order to make full use of the information, we proposed a new method, color-image-based convolutional neural networks, by converting the data into color images. Results The experiments on simulation and real data show that our method outperforms the existing methods proposed by Yue and Chen for converting data into grayscale images, in which the model accuracy is improved by an average of 7.61%, and the ratio of disease risk genes is increased by an average of 18.91%. The new method has better robustness and generalized performance.

A Semi-Supervised Fracture-Attention Model for Segmenting Tubular Objects with Improved Topological Connectivity.

Ensuring connectivity and preventing fractures in tubular object segmentation are critical for downstream analyses. Despite advancements in deep neural networks (DNNs) that have significantly improved tubular object segmentation, existing methods still face limitations. They often rely heavily on precise annotations, hindering their scalability to large-scale unlabeled image datasets. Additionally, current evaluation metrics are insufficient for effectively capturing segmentation fractures. To address these challenges, we propose a semi-supervised fracture-attention model (SSFA) for tubular object segmentation. SSFA enhances connectivity, reduces fractures, and maintains volumetric accuracy. It outperforms state-of-the-art models in topological performance. Extensive experiments on four public datasets validate the effectiveness of SSFA. Furthermore, we introduce a novel evaluation metric, the Fracture Rate (FR), which provides an intuitive and quantitative assessment of segmentation fractures. Our source code is available at http://github.com/Yanfeng-Zhou/SSFA. Supplementary data are available at Bioinformatics online.

Multi-objective optimization of urban logistics land: a gradient-based method approach with Wuhan city as an example

ABSTRACT Effective planning of logistics land is crucial for mitigating urban freight congestion, fostering economic activities, and achieving environmental equilibrium. However, the dual challenge of mismatched logistics supply and demand, along with conflicts in land use functions, can lead to inefficiencies in resource allocation and urban freight system performance. To tackle this issue, our study integrates truck GPS trajectory data with urban land use datasets to formulate a multi-objective optimization model. By utilizing the gradient descent algorithm, which effectively handles large-scale datasets, we can navigate the complexities of logistics land planning with precision. The application of this model in the Wuhan Urban Development Area reveals that: (1) across various scenarios, the model balances the utilization of multiple optimization objectives and demonstrates high solution efficiency; (2) in both the equal weight scenario and the economic preference scenario, the areas of logistics land change are characterized by high economic output, relatively good traffic conditions, greater distance from residential zones, and comparatively low land prices; and (3) based on urban development goals, the model can determine the upper bounds of the optimization objectives through manual supervision of the selection of ideal points.

A benchmark of deep learning approaches to predict lung cancer risk using national lung screening trial cohort

Deep learning (DL) methods have demonstrated remarkable effectiveness in assisting with lung cancer risk prediction tasks using computed tomography (CT) scans. However, the lack of comprehensive comparison and validation of state-of-the-art (SOTA) models in practical settings limits their clinical application. This study aims to review and analyze current SOTA deep learning models for lung cancer risk prediction (malignant-benign classification). To evaluate our model’s general performance, we selected 253 out of 467 patients from a subset of the National Lung Screening Trial (NLST) who had CT scans without contrast, which are the most commonly used, and divided them into training and test cohorts. The CT scans were preprocessed into 2D-image and 3D-volume formats according to their nodule annotations. We evaluated ten 3D and eleven 2D SOTA deep learning models, which were pretrained on large-scale general-purpose datasets (Kinetics and ImageNet) and radiological datasets (3DSeg-8, nnUnet and RadImageNet), for their lung cancer risk prediction performance. Our results showed that 3D-based deep learning models generally perform better than 2D models. On the test cohort, the best-performing 3D model achieved an AUROC of 0.86, while the best 2D model reached 0.79. The lowest AUROCs for the 3D and 2D models were 0.70 and 0.62, respectively. Furthermore, pretraining on large-scale radiological image datasets did not show the expected performance advantage over pretraining on general-purpose datasets. Both 2D and 3D deep learning models can handle lung cancer risk prediction tasks effectively, although 3D models generally have superior performance than their 2D competitors. Our findings highlight the importance of carefully selecting pretrained datasets and model architectures for lung cancer risk prediction. Overall, these results have important implications for the development and clinical integration of DL-based tools in lung cancer screening.

Homo Sapiens Chromosomal Location Ontology: A Framework for Genomic Data in Biomedical Knowledge Graphs

The Homo sapiens Chromosomal Location Ontology (HSCLO) is designed to facilitate the integration of human genomic features into biomedical knowledge graphs from releases GRCh37 and GRCh38 at multiple resolutions. HSCLO comprises two distinct versions, HSCLO37 and HSCLO38, each tailored to its respective human genome release. This ontology supports the efficient integration and analysis of human genomic data across scales ranging from entire chromosomes to individual base pairs, thereby enhancing data retrieval and interoperability within large-scale biomedical datasets. Unlike existing ontologies that primarily focus on genomic feature identification or annotation, HSCLO is specifically engineered to optimize the interoperability and scalability of genomic data within biomedical knowledge graphs. The utility and performance of HSCLO are demonstrated through a case study involving the integration of high-resolution chromatin interaction data, which reveals significant improvements in query efficiency and data linkage. HSCLO represents a valuable resource for advancing research in disease genetics, personalized medicine, and other domains that require complex genomic data integration.

Application of Google Earth Engine to Monitor Greenhouse Gases: A Review

Google Earth Engine (GEE) is a cloud-based platform revolutionizing geospatial analysis by providing access to vast satellite datasets and computational capabilities for monitoring environmental and societal issues. It incorporates machine learning (ML) techniques and algorithms as part of its tools for analyzing and processing large geospatial data. This review explores the diverse applications of GEE in monitoring and mitigating greenhouse gas emissions and uptakes. GEE is a cloud-based platform built on Google’s infrastructure for analyzing and visualizing large-scale geospatial datasets. It offers large datasets for monitoring greenhouse gas (GHG) emissions and understanding their environmental impact. By leveraging GEE’s capabilities, researchers have developed tools and algorithms to analyze remotely sensed data and accurately quantify GHG emissions and uptakes. This review examines progress and trends in GEE applications, focusing on monitoring carbon dioxide (CO2), methane (CH4), and nitrous oxide/nitrogen dioxide (N2O/NO2) emissions. It discusses the integration of GEE with different machine learning methods and the challenges and opportunities in optimizing algorithms and ensuring data interoperability. Furthermore, it highlights GEE’s role in pinpointing emission hotspots, as demonstrated in studies monitoring uptakes. By providing insights into GEE’s capabilities for precise monitoring and mapping of GHGs, this review aims to advance environmental research and decision-making processes in mitigating climate change.

Leveraging explainable AI and large-scale datasets for comprehensive classification of renal histologic types

Recently, as the number of cancer patients has increased, much research is being conducted for efficient treatment, including the use of artificial intelligence in genitourinary pathology. Recent research has focused largely on the classification of renal cell carcinoma subtypes. Nonetheless, the broader categorization of renal tissue into non-neoplastic normal tissue, benign tumor and malignant tumor remains understudied. This gap in research can primarily be attributed to the limited availability of extensive datasets including benign tumor and normal tissue in addition to specific type of renal cell carcinoma, which hampers the ability to conduct comprehensive studies in these broader categories. This research introduces a model aimed at classifying renal tissue into three primary categories: normal (non-neoplastic), benign tumor, and malignant tumor. Utilizing digital pathology while slide images (WSIs) from nephrectomy specimens of 2,535 patients from multiple institutions, the model provides a foundational approach for distinguishing these key tissue types. The study utilized a dataset of 12,223 WSIs comprising 1,300 WSIs of normal tissue, 700 WSIs of benign tumors, and 10,223 WSIs of malignant tumors. Employing the ResNet-18 architecture and a Multiple Instance Learning approach, the model demonstrated high accuracy, with F1-scores of 0.934 (CI: 0.933–0.934) for normal tissue, 0.684 (CI: 0.682–0.687) for benign tumors, and 0.878 (CI: 0.877–0.879) for malignant tumors. The overall performance was also notable, achieving a weighted average F1-score of 0.879 (CI: 0.879–0.880) and a weighted average area under the receiver operating characteristic curve of 0.969 (CI: 0.969–0.969). This model significantly aids in the swift and accurate diagnosis of renal tissue, encompassing non-neoplastic normal tissue, benign tumor, and malignant tumor.

An Enhanced Approach Using AGS Network for Skin Cancer Classification

Skin cancer accounts for over 40% of all cancer diagnoses worldwide. However, accurately diagnosing skin cancer remains challenging for dermatologists, as multiple types of skin cancer often appear visually similar. The diagnostic accuracy of dermatologists ranges between 62% and 80%. Although AI models have shown promise in assisting with skin cancer classification in various studies, obtaining the large-scale medical image datasets required for AI model training is not straightforward. To address this limitation, this study proposes the AGS network, designed to overcome the challenges of small datasets and enhance the performance of skin cancer classifiers. The AGS network integrates three key modules: Augmentation (A), GAN (G), and Segmentation (S). It was evaluated using eight deep learning classifiers—GoogLeNet, DenseNet201, ResNet50, MobileNet V3, EfficientNet B0, ViT, EfficientNet V2, and Swin Transformers—on the HAM10000 dataset. Five model configurations were also tested to assess the contribution of each module. The results showed that all eight classifiers demonstrated consistent performance improvements with the AGS network. In particular, EfficientNet V2 + AGS achieved the most significant performance gains over the baseline model, with an increase of +0.1808 in Accuracy and +0.1674 in F1-Score. Among all configurations, ResNet50+AGS achieved the best overall performance, with an Accuracy of 95.87% and an F1-Score of 95.73%. While most previous studies focused on single augmentation methods, this study demonstrates the effectiveness of combining multiple augmentation techniques within an integrated framework. The AGS network demonstrates how integrating diverse methods can improve the performance of skin cancer classification models.

Ecological dynamics of Enterobacteriaceae in the human gut microbiome across global populations.

Gut bacteria from the Enterobacteriaceae family are a major cause of opportunistic infections worldwide. Given their prevalence among healthy human gut microbiomes, interspecies interactions may play a role in modulating infection resistance. Here we uncover global ecological patterns linked to Enterobacteriaceae colonization and abundance by leveraging a large-scale dataset of 12,238 public human gut metagenomes spanning 45 countries. Machine learning analyses identified a robust gut microbiome signature associated with Enterobacteriaceae colonization status, consistent across health states and geographic locations. We classified 172 gut microbial species as co-colonizers and 135 as co-excluders, revealing a genus-wide signal of colonization resistance within Faecalibacterium and strain-specific co-colonization patterns of the underexplored Faecalimonas phoceensis. Co-exclusion is linked to functions involved in short-chain fatty acid production, iron metabolism and quorum sensing, while co-colonization is linked to greater functional diversity and metabolic resemblance to Enterobacteriaceae. Our work underscores the critical role of the intestinal environment in the colonization success of gut-associated opportunistic pathogens with implications for developing non-antibiotic therapeutic strategies.

Persistent Topological Laplacians—A Survey

Persistent topological Laplacians constitute a new class of tools in topological data analysis (TDA). They are motivated by the necessity to address challenges encountered in persistent homology when handling complex data. These Laplacians combine multiscale analysis with topological techniques to characterize the topological and geometrical features of functions and data. Their kernels fully retrieve the topological invariants of corresponding persistent homology, while their non-harmonic spectra provide supplementary information. Persistent topological Laplacians have demonstrated superior performance over persistent homology in the analysis of large-scale protein engineering datasets. In this survey, we offer a pedagogical review of persistent topological Laplacians formulated in various mathematical settings, including simplicial complexes, path complexes, flag complexes, digraphs, hypergraphs, hyperdigraphs, cellular sheaves, and N-chain complexes.

Coverage bias in small molecule machine learning

Small molecule machine learning aims to predict chemical, biochemical, or biological properties from molecular structures, with applications such as toxicity prediction, ligand binding, and pharmacokinetics. A recent trend is developing end-to-end models that avoid explicit domain knowledge. These models assume no coverage bias in training and evaluation data, meaning the data are representative of the true distribution. However, the domain of applicability is rarely considered in such models. Here, we investigate how well large-scale datasets cover the space of known biomolecular structures. For doing so, we propose a distance measure based on solving the Maximum Common Edge Subgraph (MCES) problem, which aligns well with chemical similarity. Although this method is computationally hard, we introduce an efficient approach combining Integer Linear Programming and heuristic bounds. Our findings reveal that many widely-used datasets lack uniform coverage of biomolecular structures, limiting the predictive power of models trained on them. We propose two additional methods to assess whether training datasets diverge from known molecular distributions, potentially guiding future dataset creation to improve model performance.

Default mode network functional connectivity as a transdiagnostic biomarker of cognitive function.

The default mode network (DMN) is intricately linked with processes such as self-referential thinking, episodic memory recall, goal-directed cognition, self-projection, and theory of mind. Over recent years, there has been a surge in examining its functional connectivity, particularly its relationship with frontoparietal networks (FPN) involved in top-down attention, executive function, and cognitive control. The fluidity in switching between these internal and external modes of processing-highlighted by anti-correlated functional connectivity-has been proposed as an indicator of cognitive health. Due to the ease of estimation of functional connectivity-based measures through resting state fMRI paradigms, there is now a wealth of large-scale datasets, paving the way for standardized connectivity benchmarks. This review delves into the promising role of DMN connectivity metrics as potential biomarkers of cognitive state across attention, internal mentation, mind wandering and meditation states, and investigating deviations in trait-level measures across aging and in clinical conditions such as Alzheimer's disease, Parkinson's, depression, ADHD, and others. Additionally, we tackle the issue of reliability of network estimation and functional connectivity and share recommendations for using functional connectivity measures as a biomarker of cognitive health.

FastMRI Breast: A Publicly Available Radial K-space Dataset of Breast Dynamic Contrast-enhanced MRI.

"Just Accepted" papers have undergone full peer review and have been accepted for publication in Radiology: Artificial Intelligence. This article will undergo copyediting, layout, and proof review before it is published in its final version. Please note that during production of the final copyedited article, errors may be discovered which could affect the content. The fastMRI breast dataset is the first large-scale dataset of radial k-space and DICOM data for breast dynamic contrast-enhanced MRI with case-level labels. Its public availability aims to advance fast and quantitative machine learning research. ©RSNA, 2025.