Find research datasets worth reusing
Search datasets from major research repositories and use ShareScore to quickly assess how well each record supports discovery, access, and reuse.
20
datasets available to search
ShareScore release 0.9.0
Dataset results
20 results for “variational autoencoders”
Novel multi-omics deconfounding variational autoencoders can obtain meaningful disease subtyping
<h3>TCGA pan-cancer mRNA and DNA data augmented with artificial confounders utilised in "Novel multi-omics deconfounding variational autoencoders can obtain meaningful disease subtyping" by Zuqi Li and Sonja Katz (manuscript in preparation).</h3> <p>The following data curation steps were carried out: </p> <ul> <li><strong>Step 1. Download data from TCGA</strong> <ul> <li>R package `TCGAbiolinks`</li> <li>2547 patients (after step 2) with 6 cancer types: <ul> <li>BRCA (731)</li> <li>THCA (408)</li> <li>BLCA (387)</li> <li>LUSC (297)</li> <li>HNSC (412)</li> <li>KIRC (312)</li> </ul> </li> <li>mRNA expression profiles</li> <li>DNAm expression profiles</li> <li>Clinical data: <ul> <li>tumor stage: i, ia, ib, ii, iia, iib, iii, iiia, iiib, iiic, iv, iva, ivb, ivc, x</li> <li>age at diagnosis</li> <li>race: 'white', 'black or african amarican', 'asian', 'american indian or alaska native'</li> <li>gender<br><br></li> </ul> </li> </ul> </li> <li><strong>Step 2. Removal criteria</strong> <ul> <li>Patients with <ul> <li>NA or 'not reported' clinical data</li> <li>race 'american indian or alaska native'</li> <li>tumor stage x</li> </ul> </li> <li>mRNA and DNAm probes with <ul> <li>0 variance across all included patients</li> <li>not shared across all cancer types</li> <li>with missing values<br><br></li> </ul> </li> </ul> </li> <li> <strong>Step 3. Encode clinical vairables and save datasets</strong> <ul> <li>mRNA dataset: 2547 patients x 58,456 mRNAs</li> <li>DNAm dataset: 2547 patients x 232,088 DNAm</li> <li>clinic dataset: 2547 patients x 6 variables<br> 1. patient ID<br> 2. tumor stage: 1, 1, 1, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4<br> 3. age at diagnosis<br> 4. race: asian(1), black or african amarican(2), white(3)<br> 5. gender: female(0), male(1)<br> 6. cancer type: BRCA(1), THCA(2), BLCA(3), LUSC(4), HNSC(5), KIRC(6)<br> </li> </ul> </li> <li><strong> Step 4. Pre-process the datasets</strong> <ul> <li>mRNA dataset: '<em>TCGA_mRNAs_processed.csv'</em><br> <ul> <li>Take the 2000 mRNAs with highest variance</li> <li>Rescale every feature to [0,1]</li> <li>--> 2547 patients x 2000 mRNAs</li> </ul> </li> <li>DNAm dataset: <em>'TCGA_DNAm_processed.csv'</em><br> <ul> <li>Take the 2000 DNAm with highest variance</li> <li>Rescale every feature to [0,1]</li> <li>--> 2547 patients x 2000 DNAm</li> </ul> </li> <li>clinic dataset:<em> 'TCGA_clinic.csv'<br><br></em></li> </ul> </li> <li><strong>Step 5. Simulate confounders (instructions can be found in Methods section of manuscript)</strong> <ul> <li>Linear confounder: <ul> <li><em>'TCGA_confounder_linear.csv' -</em> linear confounding classes<em><br></em></li> <li><em>'TCGA_DNAm_confounded_linear.csv' </em>- linearly confounded DNAm data<em><br></em></li> <li><em>'TCGA_mRNA2_confounded_linear.csv' </em> - linearly confounded mRNA data<em><br></em></li> </ul> </li> <li>Squared confounder <ul> <li><em>'TCGA_confounder.csv' -</em> squared confounding classes<em><br></em></li> <li><em>'TCGA_DNAm_confounded.csv' </em>- squared confounded DNAm data<em><br></em></li> <li><em>'TCGA_mRNA2_confounded.csv' </em> - squared confounded mRNA data</li> </ul> </li> <li>Categorical confounder <ul> <li><em>'TCGA_confounder_categ2.csv' -</em> categorical confounding classes<em><br></em></li> <li><em>'TCGA_DNAm_confounded_categ2.csv' </em>- categorically confounded DNAm data<em><br></em></li> <li><em>'TCGA_mRNA2_confounded_categ2.csv' </em> - categorically confounded mRNA data</li> </ul> </li> <li>Multiple confounders - combined effect (linear + squared + categorical)<br> <ul> <li><em>'TCGA_confounder_multi.csv' -</em> confounding classes for combined effect<em><br></em></li> <li><em>'TCGA_DNAm_confounded_multi.csv' </em>- DNAm data with combined effect<em><br></em></li> <li><em>'TCGA_mRNA2_confounded_multi.csv' </em> - mRNA data with combined effect</li> </ul> </li> </ul> </li> </ul> <p> </p> <p> </p>
Particle-based Fast Jet Simulation at the LHC with Variational Autoencoders: generator-level and reconstruction-level jets dataset
<p>Jets at generator and reconstruction level saved in .npy format.</p> <p>Each jet is represented as an array of jet constituents characterized by their particle momentum in Cartesian coordinates, i.e., (px, py, pz). For both generator-level and reconstruction-level jets, jet constituents are ordered by decreasing pT.</p> <p>The shape of the datasets is [N, 50, 3], where N is the total number of jets, 50 is the number of particles per jet, and 3 is the number of particle features (in order): [px, py, pz].<br> About 1.7M jets split into training, validation and testing sets at 60%, 20% and 20% respectively.</p>
Dataset for "Solving deep-learning density-functional theory via variational autoencoder"
<p>The dataset contains the ground state energies, the ground state density profiles, and the external potentials of a 3D single particle system with a Gaussian-like external potential.<br>The number of grid points for each dimension is \(N_g=18\), the linear length of the box is \(L=a_0\) with \(a_0\) the unit of length. The unit of energy is \(E_0=\frac{ \hbar^2}{(m a_0^2)}\).</p> <p> </p> <p> </p> <p>The dataset is zip file of a Python npz file with the following keys:</p> <p>- "density" that corresponds to the ground state density profile.<br>- "potential" is the external potential.<br>-"energy" is the ground state energy.</p> <p><br>The number of instances is 36000. </p> <p>-3D_gaussian.zip -> 3D_gaussian.npz</p> <p> a dictionary with three keys -density, potential, energy-.<br> The dimension of both potential and density is \([N_d,N_g,N_g,N_g]\).<br> The shape of energy is \([N_d]\).<br> \(N_d=36000\)</p> <p>-3D_gaussian_transfer_test_1.npz</p> <p> a dictionary with three keys -density, potential, energy-.<br> The dimension of both potential and density is \([N_d,N_g,N_g,N_g]\).<br> The shape of energy is \([N_d]\).<br> \(N_d=500\)</p> <p>-3D_gaussian_transfer_test_2.npz</p> <p> a dictionary with three keys -density, potential, energy-.<br> The dimension of both potential and density is \([N_d,N_g,N_g,N_g]\).<br> The shape of energy is \([N_d]\).<br> \(N_d=500\)</p>
Towards a more informative representation of the fetal-neonatal brain connectome using Variational Autoencoder
<p>Recent advances in functional magnetic resonance imaging (fMRI) have helped elucidate previously inaccessible trajectories of early-life prenatal and neonatal brain development. To date, the interpretation of fetal-neonatal fMRI data has relied on linear analytic models, akin to adult neuroimaging data. However, unlike the adult brain, the fetal and newborn brain develops extraordinarily rapidly, far outpacing any other brain development period across the lifespan. Consequently, conventional linear computational models may not adequately capture these accelerated and complex neurodevelopmental trajectories during this critical period of brain development along the prenatal-neonatal continuum. To obtain a nuanced understanding of fetal-neonatal brain development, including non-linear growth, for the first time, we developed quantitative, systems-wide representations of brain activity in a large sample (>500) of fetuses, preterm, and full-term neonates using an unsupervised deep generative model called Variational Autoencoder (VAE), a model previously shown to be superior to linear models in representing complex resting state data in healthy adults. Here, we demonstrated that non-linear brain features, i.e., latent variables, derived with the VAE pretrained on rsfMRI of human adults, carried important individual neural signatures, leading to improved representation of prenatal-neonatal brain maturational patterns and more accurate and stable age prediction in the neonate cohort compared to linear models. Using the VAE decoder, we also revealed distinct functional brain networks spanning the sensory and default mode networks. Using the VAE, we are able to reliably capture and quantify complex, non-linear fetal-neonatal functional neural connectivity. This will lay the critical foundation for detailed mapping of healthy and aberrant functional brain signatures that have their origins in fetal life.</p>
STGMVA: clustering, imputation, and integration for spatial resolved transcriptomics using spatiotemporal gaussian mixture variational autoencoder
<p> In this study, we present STGMVA, a comprehensive analysis toolkit employs a spatiotemporal gaussian mixture variational autoencoder to tackle these tasks effectively. STGMVA consists of two stages: pretraining the gene expression and spatial location using a gaussian mixture model, and learning the embedding vectors through a variational graph autoencoder. Results demonstrate STGMVA surpasses state-of-the-art approaches on various spatial transcriptomics datasets, exhibiting superior performance across different scales and resolutions. Notably, STGMVA achieves the highest clustering accuracy in human brain, mouse hippocampus, and mouse olfactory bulb tissues. Furthermore, STGMVA enhances and denoises gene expression patterns for gene imputation task. Additionally, STGMVA has the capability to correct batch effects and achieve joint analysis when integrating multiple tissue slices.</p>
Towards a more informative representation of the fetal-neonatal brain connectome using Variational Autoencoder
Open the record for dataset details and reuse information.
Data from: Bayesian estimation of muscle mechanisms and therapeutic targets using variational autoencoders
Open the record for dataset details and reuse information.
Training and test data, plus saved models for the paper "Top-down effects in an early visual cortex inspired hierarchical Variational Autoencoder" submitted to the SVRHM 2022 Workshop @ NeurIPS
<p>Each .pkl file contains a training or test dataset in the form of a Python dictionary (generated with Python 3.8.5) with the following fields:</p><ul><li>'train_images': 640,000 float32 images used for model training. 20px images contain 400 pixel intensities, 40px images contain 1600 pixel intensities each.</li><li>'train_labels': float32 labels for each image in 'train_images'. All natural images are labeled with 0.0. Texture images are labeled with 0.0, 1,0, 2.0, 3.0, or 4.0, according to their texture family.</li><li>'test_images': 64,000 float32 images used for model testing. 20px images contain 400 pixel intensities, 40px images contain 1600 pixel intensities each.</li><li>'test_labels': float32 labels for each image in 'test_images'. All natural images are labeled with 0.0. Texture images are labeled with 0.0, 1,0, 2.0, 3.0, or 4.0, according to their texture family.</li></ul><p>Each .zip file contains a saved model. Details on these are coming soon.</p><p>For more details, see the paper "Top-down effects in an early visual cortex inspired hierarchical Variational Autoencoder" published at the SVRHM 2022 Workshop @ NeurIPS (<a href="https://openreview.net/forum?id=8dfboOQfYt3">link</a>).</p>
An unsupervised deep learning framework with variational autoencoders for genome-wide DNA methylation analysis and biologic feature extraction applied to breast cancer
<p>Supplemental data for the paper titled "An unsupervised deep learning framework with variational autoencoders for genome-wide DNA methylation analysis and biologic feature extraction applied to breast cancer"</p>
Data and code from: PathVGAE: A path-based variational graph autoencoder framework for ranking centrality in road networks
Open the record for dataset details and reuse information.
Data for "MorphVAE: Generating Neural Morphologies from 3D-Walks using Variational Autoencoder with Spherical Latent Space"
<p>The reconstructions and the processed data that accompany the publication </p> <p><strong>MorphVAE: Generating Neural Morphologies from 3D-Walks using a Variational Autoencoder with Spherical Latent Space</strong></p> <p>Sophie Laturnus, Philipp Berens</p> <p>https://www.biorxiv.org/content/10.1101/2021.06.14.448271v1</p> <p>doi: https://doi.org/10.1101/2021.06.14.448271</p>
Data for "Joint Clinical and Molecular Subtyping of COPD with Variational Autoencoders"
<p>Data for paper "Data for "Joint Clinical and Molecular Subtyping of COPD with Variational Autoencoders", Maiorino et al.</p> <p>The associated code repository is at<a href="https://github.com/reemagit/joint_subtyping_vae"> https://github.com/reemagit/joint_subtyping_vae</a></p>
Unsupervised Multiple-Object Tracking with a Dynamical Variational Autoencoder
<p>This is the public dataset of synthetic trajectories and MOT17-3T, which is used in paper <a href="https://arxiv.org/abs/2202.09315">Unsupervised Multiple-Object Tracking with a Dynamical Variational Autoencoder</a>. The source code can be found <a href="https://gitlab.inria.fr/xilin/dvae-umot-release">here</a>.</p>
Dynamical Variational Autoencoders (DVAE) pretrained models
<p>The Variational Autoencoder (VAE) is a powerful deep generative model that is now extensively used to represent high-dimensional complex data via a low-dimensional latent space learned in an unsupervised manner. In the original VAE model, input data vectors are processed independently. In recent years, a series of papers have presented different extensions of the VAE to process sequential data, that not only model the latent space, but also model the temporal dependencies within a sequence of data vectors and corresponding latent vectors, relying on recurrent neural networks or state space models. In this paper we perform an extensive literature review of these models. Importantly, we introduce and discuss a general class of models called Dynamical Variational Autoencoders (DVAEs) that encompasses a large subset of these temporal VAE extensions. Then we present in detail seven different instances of DVAE that were recently proposed in the literature, with an effort to homogenize the notations and presentation lines, as well as to relate these models with existing classical temporal models. We reimplemented those seven DVAE models and we present the results of an experimental benchmark conducted on the speech analysis-resynthesis task (the PyTorch code is made publicly available). The paper is concluded with an extensive discussion on important issues concerning the DVAE class of models and future research guidelines.</p>
Multimodal omics data fusion for cancer prognosis with co-attention-based variational autoencoder
Open the record for dataset details and reuse information.
Variational Quantum Classifier-based early identification and classification of chronic kidney disease using Sparse autoencoder and Lasso Shrinkage
<p>The dataset used in this project includes patient data relevant to the identification of chronic kidney disease, including features such as:</p> <p> Age<br> Blood pressure<br> Specific blood markers (e.g., creatinine, hemoglobin)<br> Glomerular filtration rate (GFR)<br> Other medical attributes used for diagnosing CKD</p> <p>File structure of the dataset:</p> <p> chronic_kidney_disease.csv: Contains the labeled medical data for CKD classification.</p>
Dataset used in "Machine-learning Kondo physics using variational autoencoders"
<p>Spectral functions from the single-impurity Anderson model, generated on a log-linear frequency mesh using the numerical renormalization group algorithm. Used in both https://doi.org/10.1103/PhysRevB.103.245118 (where a detailed description of dataset generation can be found) and https://arxiv.org/abs/2107.08013.</p>
SpatialCVGAE: Spatial Domain Identification via Consensus Clustering Integrated Variational Graph Autoencoder
Open the record for dataset details and reuse information.
Dataset of "Towards Automatic Feature Extraction and Sample Generation of Grain Structure by Variational Autoencoder"
<p><em>No description.</em></p>
PILOT-GM-VAE: Patient-Level Analysis of Single Cell Disease Atlas with Optimal Transport of Gaussian Mixtures Variational Autoencoders
<p><strong>Datasets for PILOT-GM-VAE.<br>This part of the data includes Breast Cancer, Kidney_KPMP(only kidney tissue and all samples), Myocardial Infarction(2), and Kidney Cancer.<br>Please for more details read the data sets section of the paper.</strong></p>
ScienceDex guides
Understand access before you commit
These curated guides explain access requirements, typical timelines, costs, and reuse considerations for widely used research datasets.
Allen Brain Atlas
Allen Brain Atlas is an Allen Institute collection of brain map atlases, datasets, APIs, and analysis tools covering mouse, human, and non-human primate brain resources.
Annotated Behaviour and Observability Dataset (ABODe)
ABODe is a University of Edinburgh DataShare dataset for behavior classification in group-housed mice using home-cage video, identities, bounding boxes, ground-plate positions, and annotator labels.
DANDI Archive for NWB datasets
DANDI is a BRAIN Initiative archive for publishing and sharing neurophysiology data, including electrophysiology, optophysiology, and behavioral data packaged as NWB and related standards.
International Brain Laboratory public data
The International Brain Laboratory public data releases expose standardized mouse decision-making experiments, including Neuropixels recordings, widefield calcium imaging, behavior, and session metadata accessed through the ONE API.
OpenNeuro
OpenNeuro is a free, open platform for sharing neuroimaging datasets, with public search, dataset pages, and download paths for web, S3, DataLad, and the OpenNeuro CLI.