Find research datasets worth reusing

opencc-by-4.0Dec 2022View details →

Single-cell RNA sequencing of Sox17-expressing lineages reveals distinct gene regulatory networks and dynamic developmental trajectories

Two seurat objects contains single-cell RNA sequencing data that captures Sox17-expressing lineages during embryogenesis. sox17_integrated_Figure2B.rds : This is a seurat object that contains single-cell RNA sequencing data from integration of GFP+ cells produced from Sox17GFPCre allele marking cells that currently express Sox17 or short-term progeny of Sox17-expressing progenitors and TdTomato+ cells produced from R26LSL.TdTomato reporter allele in the presence of Sox17GFPCre marking long-term progeny of Sox17-expressing progenitors. Inferred cell types in this seurat object reflects Figure 2B in the article. sox17_Prox1_endoderm_Figure5A.rds : This is a seurat object that contain single-cell RNA sequencing data from integration of Sox17- and Prox1-expressing endoderm dataset. Prox-1 expressing endoderm data is from the Willnow et al. Nature(2021). Inferred cell types in this seurat object reflects Figure 5A in the article.

Global-Local Actions, Network and Territorial Development at Iberoamerica: The UNESCO chair in Creative Economy and Public Policy

Gravação do Painel sobre Ações Globais-Locais, Redes e Desenvolvimento Territorial na Iberoamerica: a Cátedra UNESCO em Economia Criativa e Políticas Públicas - Palestra no EnANPAD 2022 Global-Local Actions, Network and Territorial Development at Iberoamerica: The UNESCO chair in Creative Economy and Public Policy. Atividade da divisão APB no EnANPAD 2022 - <a href="http://www.anpad.org.br">www.anpad.org.br</a> Proponente/Proponent: Magnus Luiz Emmendoerfer - PPG-ADM/UFV Painelistas/Panelists: Mário Vasconcellos Sobrinho - PPAD/UNAMA and PPGEDAM/UFPA Estela Najberg - PPGADM/UFG Kelly Maria Paz e Silva - PROPAD/DCA/CCSA/UFPE Thiago Duarte Pimentel - PPGA/FACC/CMAA/UFJF and PPG/UFJF

Real world networks for network classification method evaluation

This research presents a set of network databases based on real-world data, aimed at evaluating the effectiveness of network classification methods. The first database, called the "Social database," consists of networks from the Stanford Network Analysis Project (SNAP) platform, including two classes, namely, Google+ and Twitter, we have made some preprocessment and also reduced the original dataset from the SNAP platform such that each class contains 50 networks. The other databases presented in this study are collectively known as the "Metabolic database," as they are constructed using the substrate-product network model and are based on biochemical reactions of organisms obtained from the Kyoto Encyclopedia of Genes and Genomes database (KEGG). The networks were generated using a model that considers metabolites as vertices and chemical reactions as edges. The Metabolic database comprises six classification schemes, which include the "kingdom-database" with 160 network samples, where each of the four classes contains 40 networks representing animal, plant, fungi, and protist kingdoms. The remaining databases in the Metabolic database are the "Animal-database," "Fungi-database," "Plant-database," "Firmicutes-Bacilis-database," and "Actinobacteria-database," each containing a varying number of network samples. To ensure consistency and comparability of our results, we implemented a standardization procedure whereby all networks were converted to adjacency list format. This allowed us to more efficiently process and analyze the data.

Synthetic networks for classification methods

The current study presents a collection of synthetic network databases intended for evaluating classification methods. The first database, termed "4-models," comprises synthetic networks that have been generated based on four distinct network classes, namely, Random, Small-world, Scale-free, and Geographical. This database includes 2800 networks for each class, resulting in a total of 11,200 networks. The second database, named "Scale-free," is composed of scale-free synthetic models generated using the models proposed by Barabási and Albert and by Dorogovstev and Mendes. This database includes five classes, comprising four Barabasi models based on the network's power law parameter and a Mendes model, each class consisting of 100 networks. The third database, named "Noisy," is created by modifying the topology of the networks from the 4-models and Scale-free databases via the addition or removal of edges. This database contains three sub-datasets, based on the level of randomness added to the networks, and eight different classes, each of which includes 100 networks. To ensure consistency and comparability of our results, we implemented a standardization procedure whereby all databases were converted to adjacency list format. This allowed us to more efficiently process and analyze the data.

Multivariate prediction on wake-affected wind turbines using graph neural networks (Eurodyn) database

Database consisting of graphs generated using randomized layouts and PyWake simulations used in 'Multivariate prediction on wake-affected wind turbines using graph neural networks', contribution to Eurodyn 2023.

opencc-by-4.0Dec 2020View details →

Hornet 40: Network Dataset of Geographically Placed Honeypots

Hornet 40 is a dataset of 40 days of network traffic attacks captured in cloud servers used as honeypots to help understand how geography may impact the inflow of network attacks. The honeypots are located in eight different cities: Amsterdam, London, Frankfurt, San Francisco, New York, Singapore, Toronto, Bangalore. The data was captured in April, May, and June 2021. The eight cloud servers were created and configured simultaneously following identical instructions. The network capture was performed using the Argus network monitoring tool in each cloud server. The cloud servers had only one service running (SSH on a non-standard port) and were fully dedicated as a honeypot. No honeypot software was used in this dataset. The dataset consists of eight scenarios, one for each geographically located cloud server. Each scenario contains bidirectional NetFlow files in the following format: <ul> <li>hornet40-biargus.tar.gz: all scenarios with bidirectional NetFlow files in Argus binary format;</li> <li>hornet40-netflow-v5.tar.gz: all scenarios with bidirectional NetFlow v5 files in CSV format;</li> <li>hornet40-netflow-extended.tar.gz: all scenarios with bidirectional NetFlows files in CSV format containing all features provided by Argus.</li> <li>hornet40-full.tar.gz: download all the data (biargus, NetFlow v5, and extended NetFlows)</li> </ul>

opencc-by-4.0Feb 2023View details →

Raw datasets for paper "Rapid Spatio-Temporal Flood Modelling via Hydraulics-Based Graph Neural Networks"

Raw datasets for paper "Rapid Spatio-Temporal Flood Modelling via Hydraulics-Based Graph Neural Networks". The zip folder comprises 4 subfolders (DEM, WD, VX, VY), containing the elevation, water depths in time, and velocities (in x and y directions) in time for all training and testing simulations. The overview.csv file provides the runtime of the numerical model on each different simulation, identified by its id. The simulations ids are divided as follows: - 1-80: Training and validation - 501-520: Testing dataset 1 - 10001-10020: Testing dataset 2 - 15001-15020: Testing dataset 3

Multi-scale full waveform inversion based on a convolutional neural network

The research data from this paper are uploaded here and are available for download.

Improving rainfall forecast at the district scale over the eastern Indian region using deep neural network

The code contains ANN and CNN models.

Our processed LoveDA dataset for "LOANet: A Lightweight Network Using Object Attention for Extracting Buildings and Roads from UAV Aerial Remote Sensing Images"

Our processed LoveDA dataset is used for the paper "<a href="https://doi.org/10.7717/peerj-cs.1467">LOANet: A Lightweight Network Using Object Attention for Extracting Buildings and Roads from UAV Aerial Remote Sensing Images</a>"

Our processed CITY_OSM dataset for "LOANet: A Lightweight Network Using Object Attention for Extracting Buildings and Roads from UAV Aerial Remote Sensing Images"

Our processed CITY_OSM dataset is used for the paper "<a href="https://doi.org/10.7717/peerj-cs.1467">LOANet: A Lightweight Network Using Object Attention for Extracting Buildings and Roads from UAV Aerial Remote Sensing Images</a>".

A dataset of aerial images taken by UAV that we collected for "LOANet: A Lightweight Network Using Object Attention for Extracting Buildings and Roads from UAV Aerial Remote Sensing Images"

Our private dataset of UAV aerial imagery for paper "<a href="https://doi.org/10.7717/peerj-cs.1467">LOANet: A Lightweight Network Using Object Attention for Extracting Buildings and Roads from UAV Aerial Remote Sensing Images</a>".

opencc-zeroApr 2023View details →

dryad32/100

Exploring functional flow heterogeneity in regulated flow regime: fish species turnover along hydraulic gradients in an artificial waterway network

Humans have altered river flows and lateral aquatic habitats. The expansion of agriculture in floodplains has resulted in landscapes dominated by irrigated farmland. A key challenge in water management is to conserve existing ecological communities and habitat heterogeneity, while simultaneously maintaining engineered infrastructure for agriculture. In this study, we focused on an artificial channel network for irrigation with a regulated flow regime and its function as habitat for various fish species. Differences of hydraulic conditions among channels and compositional changes in fish species were examined to clarify functional flow heterogeneity. Species turnover was analyzed using pairwise Simpson dissimilarity among sampling reaches. Species turnover was positively associated with Froude number (flow intensity) differences at intermediate discharges, and with differences in cross-sectional areas (flow magnitude) at low discharges. Drastic changes in inflows should be considered for the effective conservation of flow heterogeneity, even under a regulated flow regime. Improved engineering design to manage the hydraulic environment is one option for maintaining the ecological value of lateral waterbodies in human-dominated landscapes. Our findings provide insights into the importance of functional flow heterogeneity to conserve fish species diversity.

Double shrinking (DOSH), a regression-based algorithm for gene regulatory network inference from co-expression data

Data for the preprint "Double shrinking (DOSH), a regression-based algorithm for gene regulatory network inference from co-expression data". The preprint is live on ResearchSquare: <a href="http://t.researchsquare.com/track/click/31114617/doi.org?p=eyJzIjoiWVJQQUYtT09mMXFoWnRoMGk0SlpQZTZqWWpJIiwidiI6MSwicCI6IntcInVcIjozMTExNDYxNyxcInZcIjoxLFwidXJsXCI6XCJodHRwczpcXFwvXFxcL2RvaS5vcmdcXFwvMTAuMjEyMDNcXFwvcnMuMy5ycy0yNzM4NjgzXFxcL3YxXCIsXCJpZFwiOlwiZDMzODNjZGNhNWNiNGE2Yjk5NWRkY2UyNmIyODI5NTlcIixcInVybF9pZHNcIjpbXCIzZGQwZTAxMmExMzk4NDhkNTAzYjI4ZTBiZmU1Y2QxMDcxNzhlZTgwXCJdfSJ9">10.21203/rs.3.rs-2738683/v1</a>.

Dataset associated with Inferring Cell-Type-Specific Causal Gene Regulatory Networks during Human Neurogenesis

Full summary statistics for QTLs generated in study titled "Inferring Cell-Type-Specific Causal Gene Regulatory Networks during Human Neurogenesis" The big "data" folder includes datasets for each model under subfolders Model 1A, Model 1B and Model 2 as following tree data │   ├───Model1A │   │   ├───caQTL │   │   │   ├───neuron │   │   │   └───progenitor │   │   └───eQTL │   │       ├───neuron │   │       └───progenitor │   ├───Model1B │   └───Model2

Single-Shot Optical Neural Network

Data from classification of MNIST [1], Fashion-MNIST [2] and QuickDraw [3] images reported in "Single-Shot Optical Neural Network" by L. Bernstein et al. The networks of 784 -> N (-> N) -> 10 activations performed inference on the test sets through consecutive matrix products implemented on the optical hardware, with ReLU applied electronically between each layer (see main text for more details). Folders for each tested network contain the following text files: <ul> <li>Inputs: 2D matrices of size B x 784 containing training (B = 50,000 or 100,000), validation (B = 10,000) and test (B = 10,000) sets. Each row is an input vector that can be reshaped into an input image of size 28 x 28.</li> <li>True labels: B-length vectors containing the true label of each input in the training, validation and test sets.</li> <li>Neural network weights: 2D matrices of size K x N used in inference experiments to classify the test sets. Weights were pre-trained on the training set using a digital electronic computer as described in Materials and Methods. Weight values were normalized such that all values fall between -255 and 255. In the optical neural network, the weighting SLM displays the absolute values of the weights (rounded to the nearest integer), and the negative weight signs are applied in post-processing. The 32-bit weight values were used for inference performed on the digital electronic computer for the ground truth comparison.</li> <li>Outputs (normalized): 2D output matrices of size 10,000 x 10 from the networks processed on a digital electronic computer (ground truth) and the optical neural network. Each row is an output vector where the position of the maximum value indicates the predicted label of the input in the same row of the test set.</li> <li>Predicted labels: 10,000-element vectors that represent the labels predicted by the networks processed on a digital electronic computer (ground truth) and the optical neural network. These predicted labels were used to generate the confusion matrices and calculate the classification accuracies (versus the true labels).</li> </ul> The classes for the Fashion-MNIST dataset are the following: <ul> <li>0: T-shirt</li> <li>1: Trouser</li> <li>2: Pullover</li> <li>3: Dress</li> <li>4: Coat</li> <li>5: Sandal</li> <li>6: Shirt</li> <li>7: Sneaker</li> <li>8: Bag</li> <li>9: Ankle boot</li> </ul> And the randomly selected classes for QuickDraw are: <ul> <li>0: Hourglass</li> <li>1: Saw</li> <li>2: Golf club</li> <li>3: See saw</li> <li>4: Spoon</li> <li>5: Horse</li> <li>6: Onion</li> <li>7: Light bulb</li> <li>8: Harp</li> <li>9: Flip flops</li> </ul> [1] Y. LeCun, L. Bottou, Y. Bengio, P. Haffner, Gradient-based learning applied to document recognition. Proceedings of the IEEE 86, 2278–2324 (1998). [2] H. Xiao, K. Rasul, R. Vollgraf, Fashion-MNIST: a novel image dataset for benchmarking machine learning algorithms. Preprint at https://arxiv.org/abs/1708.07747 (2017). [3] J. Jongejan, H. Rowley, T. Kawashima, J. Kim, N. Fox-Gieg, The Quick, Draw! AI experiment, https://quickdraw.withgoogle.com/ (2016).

Uncovering Stress Fields and Defects Distributions in Graphene Using Deep Neural Networks

The trained neural networks, complete data set, and MATLAB script used to generate molecular dynamics simulation files are available here.

opennotspecifiedMar 2022View details →

FIG. 4. Haplotype network for the mitochondrial genes cytochrome b in Phylogeography of the Chocó Endemic Rainbow Characin (Teleostei: Rhoadsia)

FIG. 4. Haplotype network for the mitochondrial genes cytochrome b (Cyt-b, top) and cytochrome oxidase I (COI, bottom) color coded by site. The size of the circles is proportional to the haplotype frequency. The number of mutations between the haplotypes are represented by hatch marks. The populations within drainages are represented by different color shades (see legend). Haplotypes fall into two groups: the northern (N, enclosed by blue dashed line) and the southern group (S, enclosed by red dashed line).

Learning a quantum computer's capability using convolutional neural networks

This is supplemental data and code for: D. Hothem et al., Learning a quantum computer's capability using convolutional neural networks, (to be published). This folder contains all the data and the analysis code to generate the results presented in that paper. The core data analysis routines use PyGSTi, which can be found at <a href="https://github.com/pyGSTio/pyGSTi">https://github.com/pyGSTio/pyGSTi</a>. Please direct any questions to Daniel Hothem (dhothem@sandia.gov). NOTE: This description template was borrowed from Timothy Proctor's Zenodo entry for: Scalable Randomized Benchmarking of Quantum Computers using Mirror Circuits.

Multidimensional specialization and generalization are pervasive in soil prokaryotes with generalists dominating communities and specialists more central in networks

Project folder to run analyses in the manuscript "Multidimensional specialization and generalization are pervasive in soil prokaryotes with generalists dominating communities and specialists more central in networks" with necessary raw and processed data. Manuscript abstract: Habitat specialization underpins biological processes from species distributions to speciation. However, organisms are often described as specialists or generalists based on a single niche axis, despite facing complex, multidimensional environments. Here, we analyzed 236 prokaryotic communities across the United States demonstrating for the first time that 90% of >1,200 prokaryotes followed one of two trajectories: specialization on all niche axes (multidimensional specialization) or generalization on all axes (multidimensional generalization). We then documented that this pervasive multidimensional specialization/generalization had a wide range of ecological and evolutionary consequences. First, multidimensional specialization and generalization are highly conserved with very few transitions between these two trajectories. Second, multidimensional generalists dominated communities because they were 73 times more abundant than specialists. Lastly, multidimensional specialists played important roles in community structure with ~220% more connections in microbiome networks. These results indicate that multidimensional generalization and specialization are evolutionarily stable with multidimensional generalists supporting larger populations and multidimensional specialists playing important roles within communities likely stemming from their overrepresentation among pollutant detoxifiers and nutrient cyclers. Taken together, we demonstrate that the vast majority of soil prokaryotes are restricted to one of two multidimensional niche trajectories, multidimensional specialization or multidimensional generalization, which then has far-reaching consequences for evolutionary transitions, microbial dominance, and community roles.