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.
125
datasets available to search
ShareScore release 0.9.0
Dataset results
125 results for “Structure learning”
molxspec: Deep learning models for predicting MS2 spectra from molecular structures
<p>This repository contains a pre-processed dataset derived from the <a href="https://gnps.ucsd.edu">GNPS public repository</a> of natural product mass spectra as well as pretrained model weights for four different types of model architectures using pytorch (version 1.9.0). The contents are as follows:</p> <ul> <li>gnps_processed_data.tgz: Contains tab separated files of molecule/MS2 spectra pairs derived from GNPS after filtering for invalid structures, too large molecules (bigger than 2000 M/Z spectra), and structures that yielded valid 3D geometry optimization. The processing steps were done for positive ionization mode (pos_* files), though negative ionization data is also included (neg_* files)</li> <li>models.tgz: Contains pytorch format pretrained models for four different architecutres: MLP (a residual block multilayer perceptron trained on ECFP molecular fingerprints), BERT (the same MLP but trained on pretrained representations from the Zinc V1 pretrained ChemBERTa models on SMILES), GCN (a graph convolution architecture), and EGNN (an equivariant graph neural network). Models were trained on pos_processed_gnps_shuffled_with_3d_train.tsv found in the gnps_processed_data.tgz file described previously.</li> </ul>
Data for COLLAPSE: A representation learning framework for identification and characterization of protein structural sites
<p>Data for methods described in the paper "COLLAPSE: A representation learning framework for identification and characterization of protein structural sites" by Alexander Derry and Russ B. Altman (BioRxiv, 2022). https://www.biorxiv.org/content/10.1101/2022.07.20.500713v1</p>
Raw NGS Data for "Deep mutational scanning and machine learning reveal structural and molecular rules governing allosteric hotspots in homologous proteins"
<p>This directory contains relevant fastq files used for deep sequencing analysis in the publication “Deep mutational scanning and machine learning reveal structural and molecular rules governing allosteric hotspots in homologous proteins”. </p> <p>Fastq files are provided for presorted, uninduced and induced populations from DMS experiments of four homologs (TtgR, TetR, RolR, and MphR). Three replicates were performed for each sample.</p> <p>Data analysis of this deep sequencing data was performed using custom scripts, which are described in the methods section of the publication.</p>
Supplementary data for: Graphene Microelectrode Arrays, 4D Structured Illumination Microscopy, and a Machine Learning Spike Sorting Algorithm Permit the Analysis of Ultrastructural Neuronal Changes During Neuronal Signalling in a Model of Niemann-Pick Disease Type C
<p>Supplementary example data for the work presented in "<em>Graphene Microelectrode Arrays, 4D Structured Illumination Microscopy, and a Machine Learning Spike Sorting Algorithm Permit the Analysis of Ultrastructural Neuronal Changes During Neuronal Signalling in a Model of Niemann-Pick Disease Type C</em>". </p> <p><strong>Abstract: </strong></p> <p>Simultaneously recording network activity and ultrastructural changes of the synapse is essential for advancing our understanding of the basis of neuronal functions. However, the rapid millisecond-scale fluctuations in neuronal activity and the subtle sub-diffraction resolution changes of synaptic morphology pose significant challenges to this endeavour. Here, we use specially designed graphene microelectrode arrays (G-MEAs), which are compatible with high spatial resolution imaging across various scales as well as permit high temporal resolution electrophysiological recordings to address these challenges. Furthermore, alongside G-MEAs, we have developed an easy-to-implement machine learning algorithm to efficiently process the large datasets collected from MEA recordings. We demonstrate that the combined use of G-MEAs, machine learning (ML) spike analysis, and four-dimensional (4D) structured illumination microscopy (SIM) enables monitoring the impact of disease progression on hippocampal neurons which have been treated with an intracellular cholesterol transport inhibitor mimicking Niemann-Pick disease type C (NPC), and show that synaptic boutons, compared to untreated controls, significantly increase in size, leading to a loss in neuronal signalling capacity.</p> <p> </p>
Datasets for "Accurate and efficient structure elucidation from routine one-dimensional NMR spectra using multitask machine learning"
<p>This upload contains the datasets used for the experiments in:</p> <p>Accurate and efficient structure elucidation from routine one-dimensional NMR spectra using multitask machine learning</p> <p>Frank Hu, Michael S. Chen, Grant M. Rotskoff, Matthew W. Kanan, and Thomas E. Markland</p> <p>https://arxiv.org/abs/2408.08284</p> <p> </p> <p>For file descriptions and usage, please refer to the supplied README.md file.</p> <p> </p> <p> </p>
Multimodal learning of noncoding variant effects using genome sequence and chromatin structure
<p>ncVarPred-1D3D:</p> <p>The data used for testing the inconsistency among genome sequence, epigenetic profile, and later, to show its relation to 3D chromatin structure can be found in sanity_check_data.tar.gz.</p> <p>Some trained model for noncoding mutation effect prediction (mapping genome sequence to epigenetic profile) can be found in CNN_MLP, CNN_GCN, CNN_RNN_MLP, CNN_RNN_GCN.tar.gz.</p> <p>The trained model for pathogenic variants prediction can be found in fewshot_pathogenic_model.tar.gz. </p> <p>The training data can be found in training_data.tar.gz.</p> <p>Some noncoding variant effects prediction results, e.g. eQTL and pathogenic variants, can be replicated using the data shared in ncVar_data.tar.gz.</p>
STL files: Modeling and design of heterogeneous hierarchical bioinspired spider web structures using deep learning and additive manufacturing
<p>STL files for paper titled modeling and design of heterogeneous hierarchical bioinspired spider web structures using deep learning and additive manufacturing</p>
Multimodal learning of noncoding variant effects using genome sequence and chromatin structure
<p>ncVarPred-1D3D: pretrained models of Sei (PMID: 35817977) + our 3D structure embedding models are shared. The models are trained and validated using DeepSEA (PMID: 26301843) selected 200 bp regions (we extended to 4K bp neighboring) to predict the epigenetic profile containing 21907 epigenetic events Sei processed.</p> <p>The pretrained DeepSEA (PMID: 26301843) and reproduced DanQ (PMID: 27084946) can be found in SOTA.tar.gz.</p>
Raw Data for: "Inorganic synthesis-structure maps in zeolites with machine learning and crystallographic distances"
<p>This repository contains all the raw data to reproduce the manuscript:</p> <p>D. Schwalbe-Koda et al. "Inorganic synthesis-structure maps in zeolites with machine learning and crystallographic distances". arXiv:2307.10935 (2023)</p> <p>The raw data should be used in combination with the code hosted on GitHub: <a href="https://github.com/dskoda/Zeolites-AMD">https://github.com/dskoda/Zeolites-AMD</a>.</p> <p><strong>Description of the data</strong></p> <p>The data in this link contains all necessary information to reproduce the manuscript. In combination with the code hosted on GitHub, it can be visualized and analyzed accordingly. The full description on the columns and results is available on the GitHub code.<br> The data files in this repository are:</p> <p>- `hparams_rnd_*.json`: results of the hyperparameter optimization of all classifiers studied in this work. The data was produced by randomly sampling the train-validation-test sets. In some cases, the data was normalized (`_norm_`), and the train set was kept `balanced` or `unbalanced`.<br> - `hyp_dm`: distance matrix of all hypothetical zeolites towards the known zeolites<br> - `hyp_predictions`: predictions of the synthesis conditions for all hypothetical zeolites<br> - `xgb_ensembles*`: pickle files containing the serialized ensemble models used in the evaluation of the data in this work. The models can be loaded with the `xgboost` Python package.</p> <p><strong>License</strong></p> <p>The data and all the content from this repository is distributed under the Creative Commons Attribution 4.0 (CC-BY 4.0)</p> <p>This work was produced under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.</p> <p>Dataset released as: LLNL-MI-854709.</p>
Supplementary material for "Learning Conceptual Modeling: Structuring Overview, Research Themes and Paths for Future Research"
<p>This material containing a list of the bibliographical data of the final sample and the raw data set of 345 publications before excluding publications outside of our focus supplements the following literature review:</p> <p>Rosenthal, Kristina; Ternes, Benjamin; Strecker, Stefan, (2019). “Learning Conceptual Modeling: Structuring Overview, Research Themes and Paths for Future Research”. In: Proceedings of the 27th European Conference on Information Systems (ECIS), Stockholm and Uppsala, Sweden, June 8–14, 2019.</p>
Reconfigurations of cortical manifold structure during reward-based motor learning
Open the record for dataset details and reuse information.
DeeplyTough: Learning Structural Comparison of Protein Binding Sites (pre-processed datasets)
<p>This deposit includes the processed versions of TOUGH-M1 (Govindaraj and Brylinski, 2018), Vertex (Chen et al., 2016), and ProSPECCTs (Ehrt et al., 2018) datasets as used in our JCIM paper (Simonovsky and Meyers, 2020) - https://pubs.acs.org/doi/abs/10.1021/acs.jcim.9b00554. Specifically, this may include PDB files with associated pockets, pre-processed HTMD features, or UniProt accession number and cluster mappings. Note that these datasets have to be combined with the original dataset files released by the respective authors.</p> <p>Liability: We do not represent and/or warrant that no third party rights exist which might prevent the use of the database or that no third party rights would be infringed by said use.</p>
Data from: Automatic segmentation of multiple cardiovascular structures from cardiac computed tomography angiography images using deep learning
<p><b>Objectives: </b>To develop, demonstrate and evaluate an automated deep learning method for multiple cardiovascular structure segmentation.</p> <p><b>Background: </b>Segmentation of cardiovascular images is resource-intensive. We design an automated deep learning method for the segmentation of multiple structures from Coronary Computed Tomography Angiography (CCTA) images.</p> <p><b>Methods: </b>Images from a multicenter registry of patients that underwent clinically-indicated CCTA were used. The proximal ascending and descending aorta (PAA, DA), superior and inferior vena cavae (SVC, IVC), pulmonary artery (PA), coronary sinus (CS), right ventricular wall (RVW) and left atrial wall (LAW) were annotated as ground truth. The U-net-derived deep learning model was trained, validated and tested in a 70:20:10 split.</p> <p><b>Results: </b>The dataset comprised 206 patients, with 5.130 billion pixels. Mean age was 59.9 ± 9.4 yrs., and was 42.7% female. An overall median Dice score of 0.820 (0.782, 0.843) was achieved. Median Dice scores for PAA, DA, SVC, IVC, PA, CS, RVW and LAW were 0.969 (0.979, 0.988), 0.953 (0.955, 0.983), 0.937 (0.934, 0.965), 0.903 (0.897, 0.948), 0.775 (0.724, 0.925), 0.720 (0.642, 0.809), 0.685 (0.631, 0.761) and 0.625 (0.596, 0.749) respectively. Apart from the CS, there were no significant differences in performance between sexes or age groups.</p> <p><b>Conclusions: </b>An automated deep learning model demonstrated segmentation of multiple cardiovascular structures from CCTA images with reasonable overall accuracy when evaluated on a pixel level.</p>
Learning Context-aware Structural Representations to Predict Antigen and Antibody Binding Interfaces
<p>These are data corresponding to the paper '<a href="https://doi.org/10.1093/bioinformatics/btaa263">Learning Context-aware Structural Representations to Predict Antigen and Antibody Binding Interfaces</a>', published in Bioinformatics 2020. The code to run the experiments can be found on github <a href="https://github.com/vamships/PECAN">here</a>.</p>
Adjacency matrices and nodal attributes for prestige and homophily predict network structure for social learning of medicinal plant knowledge
<p>Human subsistence societies have thrived in environmental extremes while maintaining biodiversity through social learning of ecological knowledge, such as techniques to prepare food and medicine from local resources. However, there is limited understanding of which processes shape social learning patterns and configuration in ecological knowledge networks, or how these processes apply to resource management and biological conservation. In this study, we test the hypothesis that the prestige (rarity or exclusivity) of knowledge shapes social learning networks. In addition, we test whether people tend to select who to learn from based on prestige (knowledge or reputation), and homophily (e.g., people of the same age or gender). We used interviews to assess five types of medicinal plant knowledge and how 303 people share this knowledge across four villages in Solomon Islands. We developed exponential random graph models (ERGMs) to test whether hypothesized patterns of knowledge sharing based on prestige and homophily are more common in the observed network than in randomly simulated networks of the same size. We found that prestige predicts five hypothesized network configurations and all three hypothesized learning patterns, while homophily predicts one of three hypothesized network configurations and five of the seven hypothesized learning patterns. These results compare the strength of different prestige and homophily effects on social learning and show how cultural practices such as intermarriage can affect certain aspects of prestige and homophily. By advancing our understanding of how prestige and homophily affect ecological knowledge networks, we identify which social learning patterns have the largest effects on biocultural conservation of ecological knowledge.</p>
The datasets used in "RNA secondary structure prediction using deep learning with thermodynamic integration"
<p>The datasets used in "RNA secondary structure prediction using deep learning with thermodynamic integration"</p>
Early life learning ability predicts adult social structure, with potential implications for fitness outcomes in the wild
1. Social environments influence important ecological processes and can determine how selection acts on traits. Cognitive abilities can shape these social environments and in turn, affect individuals' fitness. 2. To understand how cognitive abilities evolve, we need to understand the complex interplay between an individual's cognitive abilities, the social environment that they inhabit and the fitness consequences of these relationships. 3. We measured the associative learning ability of pheasant chicks, Phasianus colchicus, then released them into the wild where we quantified their social position by observing their associations at feeding stations and monitored the number of days survived. 4. We observed disassortative mixing by learning performance at the population level, and poor learners had more associates than good learners. Learning was beneficial for survival when focal individuals had fewer than four associates, but survival probability across learning abilities equalized for individuals with more than four associates. 5. While the mechanisms underlying these relationships remain to be determined, the patterns of association exhibited by pheasants at feeders can be predicted by individual variation in cognitive performances and we suspect these patterns are related to differences in information use. Critically, these resulting patterns of association have fitness consequences for individuals that cannot be explained directly by their cognitive ability, but which could mediate selection on cognition.
Learning the structure of biomedical relationships from unstructured text (Part II)
<p>These files contain the seed sets and corresponding test sets of drug-gene pairs reflecting pharmacogenomic (PGx) and drug-target relationships that were used to evaluate EBC in the PLoS Comp Bio paper. Unfortunately, they didn't make it into the first upload for this paper. There are four zipped directories in this upload:</p> <p>data-drug-gene (PGx relationships, dense matrix)</p> <p>data-drug-target (drug-target relationships, dense matrix)</p> <p>data-drug-gene-fullmatrix (PGx relationships, sparse matrix)</p> <p>data-drug-target-fullmatrix (drug-target relationships, sparse matrix)</p>
Potential and training data for 'Structure-property relations of silicon oxycarbides studied using a machine learning interatomic potential'
<p>Fitted potential, training and testing data.</p>
Unleashing the power of data through organization: Structure and connections for meaning, learning, and discovery
<p><span>Knowledge organization is needed everywhere. Its importance is marked by its pervasiveness. This paper will show many areas, tasks, and functions where proper use of Knowledge Organization, construed as broadly as the term implies, provides support for learning and understanding, for sensemaking and meaning making, for inference, and for discovery by people and computer programs and thereby will make the world a better place. The paper focuses not on metadata but rather on structuring and representing the actual data or knowledge itself and argues for more communication between the largely separated KO, Ontology, Data Modeling, and Semantic Web communities to address the many problems that need better solutions. In particular, the paper discusses the application of knowledge organization in Knowledge bases for question answering and cognitive systems; Knowledge bases for information extraction from text or multimedia; Linked data; Big data and data analytics; Electronic health records as one example; Influence diagrams (causal maps), dynamic system models, process diagrams, concept maps, and other node-link diagrams; Information systems in organizations; Knowledge organization for understanding and learning; and Knowledge transfer between domains. The paper argues for moving beyond triples to a more powerful representation using entities and multi-way relationships but not attributes.</span></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.