González-González, Pablo

24 files · 776 KB · npz

Vectorial representations for common image datasets: CIFAR10, CIFAR100, CIFAR100coarse MNIST, FashionMNIST SVHN Extracted using the scripts in this repo, where more details of the extraction can be consulted. The embeddings are provided in different modalities: features: the next-to-last representation of the network logits: the pre-softmax values predictions: the post-softmax values (posterior probabilities) targets: the true labels Each file is stored in format npz (dictionary-like) and contains the numpy representations for the keys "train", "val", and "test" partitions for each dataset. The "train" partition was used to train the corresponding neural model, "val" was used for training monitoring, and "test" was not used during training. The name of the file uses the format <dataset>_<neural-model>_<modality>.npz; for example: "svhn_resnet18_logits.npz". See reference list for publications related to each dataset.

Chen, Yugeng

428 KB

This version adds the processed monthly climatological atmospheric forcing fields used for the BA and LGM FESOM experiments. The BA forcing is derived from the COSMOS 16K_0.15 experiment averaged over model years 900-1000, and the LGM forcing is derived from the COSMOS LGMW-e experiment averaged over model years 2900-3000.

Tippett, Michael

16 KB

Observed and NMME forecast Nino-3.4 SST and global mean surface temperature, 1979-2026 The observations file (obs_n34_GMSTa_197906-202605.nc) provides monthly Nino-3.4 sea surface temperature (ERSSTv5, absolute) and global mean surface temperature anomaly (NOAA GlobalTemp v6.1.0, 1991-2020 reference) for June 1979-May 2026. The NMME file (nmme_n34a_GMSTa_199106-202606.nc) provides per-model ensemble-mean Nino-3.4 SST anomaly and global mean surface temperature anomaly from seven NMME models (June initializations, 1991-2026, leads 0.5-11.5 months).

David, Cédric

50 KB

RAPID comes along with a set of test files based on a synthetic experiment called the Sandbox. More information is available in SANDBOX.md .

Zhang, Yunchong · Miller, Tim · Price, Sedona · et al.

3.8 MB

Single Sersic profile fits for 28,274 sources over Abell 2744, in all 20 NIRCam bands authors: Yunchong Zhang, Tim Miller, Sedona Price, Wren Suess, Rachel Bezanson, David Setton, and the UNCOVER/MegaScience Collaboration The sersic parameters and associated uncertainties are estimated with pysersic. The fitting targets are selected by requiring SNR>10 in the given band, using the UNCOVER/MegaScience photometric catalog. We additionally fit these observed sizes (flagged as robust or great quality) as a function of wavelength. Using the redshift posterior distribution from the UNCOVER/MegaScience SPS catalog, we infer rest-frame sizes for each available case at UV (0.3 micron), optical (0.5 micron), and NIR (1 micron), which are included in the rest-frame size catalog. See README or the catalog release paper (to be released) for details.

Altenhoff, Adrian

6 files · 100 MB · hdf5

OMAmer - tree-driven and alignment-free protein assignment to subfamilies OMAmer is an alignment-free protein family assignment method designed to avoid overly specific subfamily predictions and to scale efficiently to phylogenomic databases containing thousands of genomes. It relies on an innovative approach that uses evolutionarily informed k-mers for alignment-free mapping to ancestral protein subfamilies. This dataset provides precomputed OMAmer databases derived from the Hierarchical Orthologous Groups in the OMA Browser . We aim to update these databases with every new OMA Browser release. Each OMAmer database is built using the latest version of the OMAmer package available at the time of the corresponding OMA Browser release. The dataset includes databases for different subsets of the species taxonomy. In most cases, we recommend using the LUCA.h5 database, which contains information from all species in the OMA database. The subset-specific databases are mainly useful when disk space is limited. The release May2026 is based on the OMA Browser release May 2026 which comprises 2983 species. We used OMAmer version 2.1.0 to build these databases.

Meng, Wen · Das, Arun · Sinha, Harsh · et al.

8.0 MB

Kaposi's sarcoma (KS) is a highly inflammatory, angiogenic tumor driven by Kaposi's sarcoma-associated herpesvirus (KSHV), yet the origins of tumor cells and mechanisms of progression remain unclear. Here, we present the first spatial single-cell atlas of KS, profiling 256 samples across patch, plaque, and nodular lesions and normal controls. We identify CD34⁺ progenitor lymphatic endothelial cells (LECs) as the primary targets of KSHV, whose clonal expansion drives tumor growth. KSHV infection induces widespread cellular reprogramming across the tumor microenvironment, including LECs, vascular endothelial cells, fibroblasts, and macrophages, generating hybrid phenotypes that support angiogenesis, inflammation, and immune modulation. KSHV⁺ macrophages are enriched in tumor-proximal niches, further promoting a proangiogenic, immunosuppressive environment. Spatial analysis reveals evolving tumor-associated niches, with a core-to-periphery gradient correlating with infection, immune modulation, and cellular remodeling. We identify disease progression predictive signatures, offering mechanistic insights into KS pathogenesis and potential new therapeutic strategies by reprogramming the tumor microenvironment.

Bartels, Paige · Oyola-Merced, Mayra · L'Ecuyer, Tristan

100 MB

The synergistic collocation of the PREFIRE, MERRA-2 Reanalysis, and EarthCARE datasets provides novel opportunities to analyze the radiative effects of clouds and aerosols in the Arctic. This dataset has a spatial resolution of approximately 1.4 km x 1.7 km, which is the same as the spatial resolution of EarthCARE's CPR. This dataset also contains the inputs or the requirements to derive inputs for a radiative transfer model. ESA/JAXA EarthCARE launched in 2024 with the mission to improve climate and weather forecasts. Of the four instruments on board, the 355 nm Atmospheric Lidar (ATLID) and 94 GHz Cloud Profile Radar (CPR) provide this dataset with extinction, backscatter, and depolarization (EBD) and nominal 1B data, respectively. NASA PREFIRE CubeSats were also launched in 2024, intending to measure far-infrared (FIR) radiation at the poles to reduce uncertainty of Arctic clouds in the global radiation budget. Each of the two CubeSats provides pushbroom longwave spectral radiometry measurements, ranging from 5 to 54 microns with a resolution of approximately 0.84 microns. Finally, the NASA GMAO MERRA-2 Reanalysis provides hourly global gridded aerosol-considered thermodynamic and aerosol mixing ratio data. At the time of publishing, MERRA-2 does not assimilate EarthCARE nor PREFIRE data. For more methodology and variable information, consult the README.md document. The data files are in a .nc4 format, with the following naming conventions: arctic_earthcare_prefire_merra2_YYYYMMDDTHHmmSS_vN, where YYYY is the four-digit year, MM is the two-digit month, DD is the two-digit day, HH is the two-digit hour, mm is the two-digit minute, SS is the two-digit seconds, and N is the version number. The datetimes are taken from the middle of the EarthCARE/PREFIRE collocation period. Source data acknowledgments: This dataset contains collocated and processed data derived from PREFIRE, EarthCARE, and MERRA-2 products. Original datasets are provided by NASA and ESA/JAXA and remain subject to their respective data-use policies.

Ng, Chi Fung Joseph

8.0 MB

To simulate the situation of having new data to annotate & integrate, we use here a scRNA-seq study ( https://pmc.ncbi.nlm.nih.gov/articles/PMC7439502/ ) of lung cells from idiopathic pulmonary fibrosis (IPF), a fatal interstitial lung disease. The raw data can be found here on the GEO database ( https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE136831 ). The .h5ad data object here was generated in the following way: .gz files were downloaded from GSE136831 (mtx of transcript counts, cell metadata, gene and cell identifiers) and assembled into an AnnData object (n_obs × n_vars = 312928 × 45947). The following cell types (from metadata: "Manuscript_Identity" column) were retained: [ "Macrophage", "Macrophage_Alveolar", "cMonocyte", "ncMonocyte", "Fibroblast", "Myofibroblast", "Aberrant_Basaloid"] The "Macrophage", "Macrophage_Alveolar", "cMonocyte" and "ncMonocyte" cell types were overrepresented relative to the other cell-types retained. We therefore downsampled specifically these 4 cell types to 5000 cells each. All cells labelled as the other retained cell-types were retained. The cells from IPF samples (metadata column "Disease_Identity" == "IPF") were further subsetted to leave n = 12651 cells.

Novelli, Chiara

8.0 MB

This dataset provides a 2D snapshot matrix generated from numerical simulations of the combustion of hydrogen with flashback. The matrix contains temporally evolving fields of: Density (&rho;) Hydrogen mass fraction (YH2) Water vapor mass fraction (YH2O) Hydroxyl radical mass fraction (YOH) Temperature (T) Data structure : Spatial grid: nx=601 points in x&isin;[0,0.06]&thinsp;m ; ny=201 points in y&isin;[-0.01,0.01]&thinsp;m. Features are concatenated into a single snapshot matrix X with dimensions (nspace,ntime) , where n_space = n_features × nx × ny. Stored as NumPy binary file ( .npy ).

Lu, Bin · Xin, Yi · Jin, Meng

8.0 MB

Jingwei-DO is a deep learning-based global ocean dissolved oxygen (DO) mapping dataset. The model is trained with quality-controlled dissolved oxygen observations from the World Ocean Database 2023 (WOD23) updated through 2025, and is applied to EN4.2.2 temperature and salinity fields for global inference. The product provides monthly dissolved oxygen fields on a 1° × 1° global grid from the surface to 5,500 m depth over the period 1965-2025. An open visualization and exploration platform is available at https://jingwei.acemap.info/ .

Nicola, Mayer · Alexander, Löhr · Nimrod, Moiseyev · et al.

8.0 MB

Data for "Enantiosensitive exceptional points in open chiral systems" This repository contains the data files, Jupyter notebooks and Gaussian16 input and output files needed to reproduce the numerical results from the paper "Enantiosensitive exceptional points in open chiral systems" , N. Mayer, A. Löhr, N. Moiseyev, M. Ivanov and O. Smirnova, Physical Review A (2026), doi.org/10.1103/7rtp-3xts arXiv:2502.18963 as well as the results contained in its Supplementary Files. Please refer to the paper for a detailed description of the numerical approach behind the numerical simulations. Note that the figures in the paper are assembled in a separate software from the plots reproduced here. For a detailed description of the files, see README.odt. The copyright of this collection rests with the authors (2026). It is made available under the Creative Commons Attribution-ShareAlike 4.0 ( CC BY-SA 4.0 ) license. For any academic use that results in a publication, please cite the main paper in addition to this deposit.

Cai

12 files · 8.0 MB · npy

flores rojas, Jose luis

8.0 MB

Title Quality-controlled gradient-tower meteorological profiles and surface-flux estimates from the Huancayo Geophysical Observatory, Peruvian central Andes Alternative short title HYGO gradient-tower surface-flux dataset Resource type Dataset Version v1.0 Creators Flores-Rojas, José Luis; Pérez Tello, María; Fashé-Raymundo, Octavio; Pareja Quispe, David; Eche Llenque, José Carlos; Silva, Yamina; Zuñiga Huaman, Gerson Description This dataset contains quality-controlled gradient-tower meteorological profiles and surface-flux estimates from the Huancayo Geophysical Observatory (HYGO) of the Geophysical Institute of Peru (IGP), located in the Mantaro Valley of the Peruvian central Andes. HYGO is a high-altitude agricultural and atmospheric observatory representative of complex Andean terrain, strong diurnal forcing, seasonal moisture contrasts, and mountain-valley circulations. The dataset was developed to support reproducible analysis of near-surface atmospheric structure and turbulent exchange in complex terrain. Native 1-min observations of air temperature, relative humidity, wind speed, and wind direction were processed through a documented workflow that includes timestamp auditing, primary meteorological quality control, conservative bit-mask flagging, thermodynamic derivation, 30-min aggregation, flux-specific pre-calculation quality control, dual-method turbulent-flux estimation, method-status diagnostics, and post-calculation plausibility filtering. The released products include cleaned 1-min tower observations, per-sample QC flags, derived thermodynamic variables, 30-min aggregated profiles, and surface-flux estimates obtained with two aerodynamic approaches: Monin-Obukhov Similarity Theory (MOST) and an anchored multi-layer Bulk Richardson Number method (BRN_ANC). The flux products include friction velocity, sensible heat flux, latent heat flux, Obukhov length, bulk Richardson number, method-status flags, post-calculation QC flags, raw method outputs, and post-QC-filtered outputs. This structure allows users to distinguish between input-profile limitations, numerical method failures, and physically implausible flux estimates. The gradient-tower system includes measurements at multiple levels between 2 and 29 m above ground level. For the flux-gradient calculations, the 2, 6, 12, and 24 m levels were used to construct vertical profiles of wind speed, temperature, humidity, and virtual potential temperature. The 18 and 29 m levels were used for wind-direction information where available but were not included in the main flux-gradient calculations. The dataset is intended for boundary-layer research, land-atmosphere interaction studies, evaluation of surface-layer parameterizations, comparison of MOST and Richardson-number methods, model validation, agricultural micrometeorology, frost-risk assessment, drought-related studies, and development of reproducible workflows for high-frequency meteorological tower data. Dataset period [Insert final period, e.g., 15 May 2018 to 30 April 2026] Geographic coverage Huancayo Geophysical Observatory, Mantaro Valley, central Peruvian Andes Latitude: [insert final latitude, e.g., -12.04145] Longitude: [insert final longitude, e.g., -75.31875] Elevation: [insert final elevation, e.g., 3315 m a.s.l.] Temporal resolution 1 min for native and cleaned meteorological observations. 30 min for aggregated profiles and turbulent-flux products. Main variables Air temperature Relative humidity Wind speed Wind direction Atmospheric pressure Saturation vapour pressure Actual vapour pressure Water-vapour mixing ratio Specific humidity Virtual potential temperature Friction velocity Sensible heat flux Latent heat flux Obukhov length Bulk Richardson number Input QC flags Method-status flags Post-calculation QC flags 30-min availability diagnostics Processing summary Raw 1-min tower observations were time-sorted, audited for duplicate timestamps, and regularized to a 1-min temporal grid when required. A primary meteorological QC system generated per-sample bit-mask flags for missing values, range violations, step changes, persistence, spikes, calm wind, humidity inconsistency, and resample-inserted timestamps. Hard-fail values were removed under a conservative rule: RANGE or simultaneous STEP and SPIKE. Contextual flags were retained for diagnostic use. Thermodynamic variables were derived after QC, including vapour-pressure variables, specific humidity, and virtual potential temperature. Cleaned 1-min profiles were aggregated to 30-min profiles with availability diagnostics. Flux-specific pre-calculation QC screened each 30-min profile before flux estimation. MOST and BRN_ANC flux estimates were computed independently from the same eligible profiles. Method-status flags recorded numerical success, non-convergence, invalid profile slopes, Richardson-number exceedance, and other execution outcomes. Post-calculation QC retained physically plausible flux estimates and masked non-passing values in the final filtered output columns. Raw method outputs were preserved separately to support diagnostic audits and sensitivity analyses. File contents [Edit this list to match the final Zenodo upload.] cleaned_1min_tower_data.[nc/csv] Cleaned 1-min meteorological observations and primary QC flags. derived_thermodynamic_variables.[nc/csv] Pressure, vapour-pressure variables, mixing ratio, specific humidity, and virtual potential temperature. aggregated_30min_profiles.[nc/csv] Thirty-minute mean profiles and data-availability diagnostics. surface_fluxes_MOST_BRN_ANC_30min.[nc/csv] MOST and BRN_ANC flux estimates, method-status flags, post-QC flags, raw outputs, and filtered outputs. qc_flag_dictionary.[csv/json] Definitions of primary QC bit masks, pre-calculation QC flags, post-calculation QC flags, and method-status flags. processing_scripts.[zip] Python scripts used for QC, thermodynamic derivation, aggregation, MOST, BRN_ANC, post-QC, diagnostics, and figures. environment.[yml/txt] Software environment and package dependencies required to reproduce the workflow. README.md Dataset description, file structure, variable names, units, QC interpretation, and recommended use. Recommended citation Flores-Rojas, J. L., Pérez Tello, M., Fashé-Raymundo, O., Pareja Quispe, D., Eche Llenque, L. Suárez Salas, J. C., Silva, Y., and Zuñiga Huaman, G. ([year]). Quality-controlled gradient-tower meteorological profiles and surface-flux estimates from the Huancayo Geophysical Observatory, Peruvian central Andes (Version v1.0) Keywords gradient tower; surface energy fluxes; quality control; Monin-Obukhov Similarity Theory; MOST; Bulk Richardson number; BRN_ANC; atmospheric surface layer; boundary layer; turbulent fluxes; sensible heat flux; latent heat flux; friction velocity; tropical Andes; Mantaro Valley; Huancayo Geophysical Observatory; HYGO; Peru; micrometeorology; land-atmosphere interactions; reproducible workflow License [Recommended: Creative Commons Attribution 4.0 International, CC BY 4.0, if allowed by your institution and funder.] Related identifiers Is supplement to: [insert article DOI after publication] Is documented by: [insert manuscript/preprint DOI if available] Is supplemented by: [insert software DOI if scripts are archived separately] Is version of: [insert previous Zenodo DOI if this is an updated version] Funding Instituto Geofísico del Perú; PROCIENCIA project "Fortalecimiento del Laboratorio de Microfísica Atmosférica y Radiación para el estudio de la interacción superficie-atmósfera en una zona agrícola de los Andes Centrales del Perú, en el contexto de cambio climático" (LAMAR), Contract No. PE501086050-2023-PROCIENCIA-BM. Notes Users should treat the flux estimates as gradient-based products, not as direct eddy-covariance measurements. MOST and BRN_ANC estimates are provided together to support method comparison and uncertainty assessment. Strongly stable, weak-wind, transition-period, and horizontally heterogeneous conditions may increase uncertainty. Users are encouraged to use the QC flags, method-status flags, and raw-output variables when performing sensitivity analyses or applying stricter filters.