Find research datasets worth reusing

Predicted room temperature electrical conductivity of molecular mixtures

In the associated manuscript, we propose the MolSets machine learning model for molecular mixture properties. Using the MolSets architecture, we train a model on a dataset curated by Bradford et al. (2023) to predict the room temperature (298 K) electrical conductivity of mixtures. Here, we report the model-predicted conductivities of all equal-weight binary mixtures among 28 types of small molecules, combined with 30 types of Li+ salts (1 mol·kg-1), totaling 11,340 candidate lithium battery electrolytes. Note that the current model has a limitation of not taking salt solubility into account. This dataset is for demonstration purposes and should be used with caution.

The more the better: Fatty acids are predictive markers of honey bee, Apis mellifera, worker longevity

Fatty acids (FA), stemming from nutrition, form triglycerides that are key components for insect energy reserves. In managed Apis mellifera colonies, supplementary feeding is common practice, yet micronutrients and microbiota (i.e. B-vitamins and probiotics) are often neglected. Given that B-vitamins are obligate cofactors for FA metabolism, and probiotics likely play key roles as well (i.e. Lactobacillus spp. synthesize B-vitamins), understanding how they contribute to FA acquisition remains unknown. Indeed, FAs are established predictors to A. mellifera longevity, and as such, are a logical point of interest in long-lived "winter" bees, where A. mellifera colony losses typically occur. Here, in a hoarding cage trial, freshly emerged adult winter workers were exposed to antibiotics (ABX) to decouple innate benefits associated to native gut microbiota, or left unexposed to ABX (N=72 cages, N=2088 experimental workers). Subsequently, all workers were fed different diets containing either probiotics, B-vitamins, with replicate treatments given ad libitum access to pollen (mimicking real-hive scenarios) or left blank (control). At the end of the trial, a subsample (n=356) had their total FA contents analyzed using Gas Chromatography coupled to Flame Ionization Detector (GC-FID). Irrespective of dietary treatment, every worker contained all 11 identified FAs, aligning our results with a priori evidence and highlighting their underlying key roles for bee physiology and health. We show for the first time that B-vitamins alone did not improve the overall abundance of individual FAs (g), yet significant differences were associated with presence/absence of bacteria and/or access to pollen, reconfirming likely ties of microbiota aiding in nutrient breakdown of complex polysaccharides found in pollen. Finally, of clear importance, there was a positive significant correlation between total lipid content and worker longevity (+2.4 median day lifespan increase / mg of FA), thereby confirming the relevance of FAs for honey bee worker longevity.

Fast-slow traits predict competition network structure and its response to resources and enemies

Plants interact in complex networks but how network structure depends on resources, natural enemies, and species resource-use strategy remains poorly understood. Here, we quantified competition networks among 18 plants varying in fast-slow strategy, by testing how increased nutrient availability and reduced foliar pathogens affected intra- and inter-specific interactions. Our results show that nitrogen and pathogens altered several aspects of network structure, often in unexpected ways due to fast and slow-growing species responding differently. Nitrogen addition increased competition asymmetry in slow-growing networks, as expected, but decreased it in fast-growing networks. Pathogen reduction made networks more even and less skewed because pathogens targeted weaker competitors. Surprisingly, pathogens and nitrogen dampened each other's effect. Our results show that plant growth strategy is key to understanding how competition responds to resources and enemies, a prediction from classic theories that has rarely been tested by linking functional traits to competition networks.

Dataset of management and growth curve derivatives to predict first lactation milk yield

This dataset presents management and modeled variables of 78 Holstein-Friesian heifers. Spreadsheets consist of seven sets of training and testing data randomly split with replacement. In each spreadsheet, the included variables are: <table> <tbody> <tr> <td>Variable name</td> <td>Description</td> </tr> <tr> <td>ID</td> <td>Heifer ID tag</td> </tr> <tr> <td>Mes</td> <td>Month of birth of the heifer</td> </tr> <tr> <td>Estacion</td> <td>Birth season, according to Mes</td> </tr> <tr> <td>L_305_F_2x2</td> <td>First lactation milk yield is calculated as the integral from day 1 to 305 of a Fourier model representing daily milk yield (kg)</td> </tr> <tr> <td>Edad_IA</td> <td>Heifer's age at effective artificial insemination</td> </tr> <tr> <td>servicios</td> <td>Number of inseminations</td> </tr> <tr> <td>Edad_1P</td> <td>Heifer's age at first calving</td> </tr> <tr> <td>P_parto</td> <td>Heifer's weight at first calving (kg)</td> </tr> <tr> <td>B_parto</td> <td>Heifer's condition score at first calving</td> </tr> <tr> <td>Sexo_C</td> <td>Calf sex</td> </tr> <tr> <td>Peso_C</td> <td>Calf weight at birth (kg)</td> </tr> <tr> <td>wt_3._F, wt_6._F, wt_9._F, wt_12._F, wt_15._F, wt_18._F, wt_21._F</td> <td>Heifer's weight at various months after birth, as per a Fourier model of Heifer's growth curve (kg)</td> </tr> <tr> <td>X1STd_3._F, X1STd_6._F, X1STd_9._F, X1STd_12._F, X1STd_15._F, X1STd_18._F, X1STd_21._F</td> <td>First derivative at several months after birth, based on a Fourier model of Heifer's growth curve</td> </tr> <tr> <td>X2STd_3._F, X2STd_6._F, X2STd_9._F, X2STd_12._F, X2STd_15._F, X2STd_18._F, X2STd_21._F</td> <td>Second derivative at several months after birth, based on a Fourier model of Heifer's growth curve</td> </tr> <tr> <td> wt_3._L to wt_21._L, X1STd_3._L to X1STd_21._L,  X2STd_3._P to X2STd_21._P </td> <td>Variables generated from a Power model</td> </tr> </tbody> </table>

QCxMS prediction of alkyl halides comparison of GFN1-xTB and GFN2-xTB

Exported Galaxy history containing spectral matching results to compare QCxMS predictions of alkyl halides using GFN1-xTB and GFN2-xTB against their high resolution experimental spectra.

Baseline glucocorticoids alone do not predict reproductive success across years but in interaction with enzymatic antioxidants

Glucocorticoids are known to adjust organismal functions, such as metabolism, in response to environmental conditions. Therefore, these hormones are thought to play a key role in regulating the metabolically demanding aspects of reproduction, especially in variable environments. However, support for the hypothesis that variation in glucocorticoid concentrations predicts reproductive success is decidedly mixed. Two explanations may account for this discrepancy: a) glucocorticoids might not act independently but could interact with other physiological traits, jointly influencing reproduction, and b) such an association could become apparent primarily in challenging environments when glucocorticoid concentrations increase. To address these two possibilities, we determined natural variation in circulating baseline glucocorticoid concentrations in parental great tits (Parus major) alongside two physiological systems known to be related to an individual's metabolism: oxidative status parameters (i.e., concentrations of pro-oxidants, dietary, and enzymatic antioxidants) and body condition. These systems interact with glucocorticoids and can also influence reproductive success. We measured these variables in two breeding seasons that differed in environmental conditions. When accounting for the interaction of baseline glucocorticoids with other physiological traits, we found a positive relationship between baseline glucocorticoids and the number of fledglings in adult great tits. The strength of this relationship was more pronounced for those individuals who also had high concentrations of the enzymatic antioxidant glutathione peroxidase. When studied independently, glucocorticoids were not related to fitness proxies, even in the year with more challenging environmental conditions. Together, our study lends to support the hypothesis that glucocorticoids do not influence fitness alone, but in association with other physiological systems

Train and Evaluation Code, Road Classification Models and Test set of the paper "Impact of Image Resolution and Image Overlap on the Prediction Performance of Convolutional Neural Networks Trained for Road Classification"

This repository contains the Python scripts built for training and evaluation of the implementation, together with the test data and the resulting road classification models corresponding to the paper "Impact of Image Resolution and Image Overlap on the Prediction Performance of Convolutional Neural Networks Trained for Road Classification". The scripts make use of the Tensorflow with Keras framework and the additional required dependencies. The training and validation set is based on the binary SROADEX dataset (https://zenodo.org/records/6482346) that was re-split into tiles that feature the image resolutions (256 x 256, 512 x 512, and 1024 x 1024 pixels) and image overlaps (0% and 12.5%) considered in this study. The data have been generated using scripts developed in Python using Open Source libraries (GDAL/OGR and MapScript) for rasterization of vector cartography that represents the axes of the different types of roads (urban, interurban and rural). This binary road data contains information from 16 full orthoimages (28.5 km * 18.5 km) with spatial resolution of 0.5 m/pixel from the insular and peninsular Spanish territory. Due to the size on disk of approximately 546 gigabytes, this training and validation data is only available upon request from the corresponding author. The test set has been generated from a novel area of 28.5 km * 18.5 km and features binary road labels. The test sets are provided in the repository for each resolution (with no overlap), so that additional DL models can be evaluated on the same data and compared with the results achieved in this study. The structure of the information shared in this repository is as follows: The scripts have been grouped by tile resolution (256, 512 and 1024). First, the test set and the evaluation script can be found. For each tile resolution, there are two subfolders (corresponding to the "no overlap" and "12.5% overlap"). In each case, the Python scripts for training the models in the three repetitions are shared, and the trained models (H5 format) are shared in compressed form. Finally, for each resolution we also share the testing dataset which consists of two folders. The material is distributed under a CC-BY 4.0 license.

Wind and SOLAR RES predicted production data for Crete and Peloponnese - ONENET WP8

WP8 aimed at the development and implementation of a web based app that enhances Active Power Management necessary for coordination of a TSOs and DSOs, using AI methods and cloud calculation engines that was tested in Peloponnese and Crete regions. Full description of the scope and results of WP8 Greek demo can be found in the relative deliverable D8.2: Development and implementation of the “F-Channel” platform (https://www.onenet-project.eu//wp-content/uploads/2023/10/OneNet_D8.2_V1.0.pdf). For purpose of this project similar, historical weather data in 1 hour resolution have been used in order to obtain behavior patterns of climatic parameters (daily, monthly, season) throughout region of interest. For this purpose various ERA5 climatic datasets has been used and AI algorithms applied in combination with terrain orography data. Modeled results was compared with operational data from TSO/DSO and appropriate model calibration has been provided based on deep learning AI algorithms. The data for Wind power plant modelled production is given in file <<a href="../api/records/10848817/draft/files/wind_res_onenet_wp8.csv/content" target="_blank" rel="noopener noreferrer">wind_res_onenet_wp8.csv</a>> in the following columns <time> ; <aa> ; <pw> ; <ws> ; <wp_name> . Columns are related to: Hourly time, wind park code, power [MWh], wind speed [m/s] and wind park name, respectively. The data for Solar power plant modelled production is given in file <<a href="../api/records/10848817/draft/files/solar_onenet_wp8.csv/content" target="_blank" rel="noopener noreferrer">solar_onenet_wp8.csv</a>> in the following columns <time> ; <name> ; <pw> ; <ta> ; <ghi> . Columns are related to: Hourly time, solar park name, power in MWh, ambient temperature and Global Horizontal irradiance [W/m2], respectively.

opencc-by-4.0Oct 2023View details →

Replication Data for the paper "Predicting Food-Security Crises in the Horn of Africa Using Machine Learning"

This folder contains all input data necessary to run the machine learning model as described in the paper "Predicting Food Security Crises in the Horn of Africa Using Machine Learning".  This model is developed at the Institute for Environmental Studies, Vrije Universiteit Amsterdam. Questions or remarks can be send to tim.busker@vu.nl

Predictions of critical events on OHLs - EUROPE ONENET WP8

WP8 aimed at the development and implementation of a web based app that enhances Active Power Management necessary for coordination of a TSOs and DSOs, using AI methods and cloud calculation engines that was tested in Peloponnese and Crete regions. Full description of the scope and results of WP8 Greek demo can be found in the relative deliverable D8.2: Development and implementation of the “F-Channel” platform (https://www.onenet-project.eu//wp-content/uploads/2023/10/OneNet_D8.2_V1.0.pdf). For critical OHL events, results of single run of weather forecast along  predetermined line route return critical events are described in following. For OHL extreme icing events data is given in file : <<a href="../api/records/10849033/draft/files/critical_ohl_events_onenet_wp8_icing.csv/content" target="_blank" rel="noopener noreferrer">critical_ohl_events_onenet_wp8_icing.csv</a>>  in following columns: <d1> ; <d2> ; <t2m> ; <req> ; <name> ; <operator> ; < voltage> ; <near_poi_name> ; <reg1> ; <reg2> ; <cntr_code >  Where columns represent: Start time of weather event, End time of weather event, temperature [deg C], OHL accumulated ice thickness [mm], name of line, operator of line, voltage of line and geographical references for affected area, respectively.   For OHL extreme wind and precepitation load events data is given in file : <<a href="../api/records/10849033/draft/files/critical_ohl_events_onenet_wp8_wind_prec.csv/content" target="_blank" rel="noopener noreferrer">critical_ohl_events_onenet_wp8_wind_prec.csv</a>>  in following columns: <d>  ; <t2m> ; <wavg> ; <tp> ;<name> ; <operator> ; < voltage> ; <near_poi_name> ; <reg1> ; <reg2> ; <cntr_code >  Where columns represent: Time  of weather event, temperature [deg C], average wind speed [m/s], precepitation rate [mm/h], name of line, operator of line, voltage of line and geographical references for affected area, respectively.   For OHL extreme wind load events data is given in file : <<a href="../api/records/10849033/draft/files/critical_ohl_events_onenet_wp8_wind.csv/content" target="_blank" rel="noopener noreferrer">critical_ohl_events_onenet_wp8_wind.csv</a>>  in following columns: <d>  ; <t2m> ; <wmax> ;<name> ; <operator> ; < voltage> ; <near_poi_name> ; <reg1> ; <reg2> ; <cntr_code >  Where columns represent: Time  of weather event, temperature [deg C], maximum wind speed [m/s], name of line, operator of line, voltage of line and geographical references for affected area, respectively.

Data for: Predictable local adaptation in butterfly photoperiodism but not thermal performance along a latitudinal cline

In seasonal environments, organisms must synchronize their life cycles to conditions favorable for growth and reproduction. Because season length varies geographically, local adaptation should arise in traits that regulate phenological responses. Geographic photoperiodism clines are well-known, but comparable studies on thermal performance are equivocal and often overlook non-linear responses. Therefore, we examined local adaptation in plastic responses to both photoperiod and temperature along a 752 km latitudinal cline, by comparing four Swedish populations of the butterfly Pieris napi. Using a common garden design, we estimated (1) photoperiod response curves for diapause induction and (2) thermal performance curves for development and growth rates. We show that differences in photoperiodism follow the expected geographical pattern, where diapause is induced at longer daylengths in northern populations (where growth seasons are short and summer days long). However, population differences in thermal performance curves were small and seemingly idiosyncratic, without clear clinal patterns. Photoperiodic responses appear to evolve more readily than thermal responses, highlighting photoperiodism as a key driver of local life cycle synchronization.

Simulated Herbarium data for testing the accuracy with which specimen data can predict the timing and duration of population-level flowering displays

This dataset provides code and example data for simulating specimen collections of flowering plants across North America, and for developing phenological predictions of population-level flowering onset and termination for these data. It further presents code for assessing the accuracy of these predictions relaticve to known (simulated) population-level flowering dates at the location of each collection.

Groundwater level data used in the manuscript titled "An explainable Bayesian TimesNet for probabilistic groundwater level prediction with application to semi-arid regions"

The standardized semimonthly groundwater levels collected from 30 monitoring wells in Dalad County, China. The  data were obtained from the Ministry of Water Resources of China and the groundwater yearbooks.  The data are used in our submitted manuscript titled "An explainable Bayesian TimesNet for probabilistic groundwater level prediction with application to semi-arid regions". If you find this dataset useful for your research, please consider to cite our manuscript upon publication.

MESSENGER_FNN_IMF_Predictions_27_03_2024

This dataset contains MESSENGER magnetosheath measurements and predictions made by a feed-forward neural network. The parameters are in the aberrated Mercury magnetospheric coordinate system (MSM'). The time cadence for these measurements is 40 s. The columns are as follows: <ol> <li>magx: x component of the magnetosheath magnetic field (nT)</li> <li>magy: y component of the magnetosheath magnetic field (nT)</li> <li>magz: z component of the magnetosheath magnetic field (nT)</li> <li>magamp: amplitude of the magnetosheath magnetic field (nT)</li> <li>time: time of magnetosheath measurement (YEAR-MONTH-DAY HOUR:MINUTE:SECOND)</li> <li>x: position of spacecraft along the X_MSM' axis</li> <li>r: cylindrical radial position (sqrt(Y_MSM'^2+Z_MSM'^2))</li> <li>theta: angular position tan^(-1)(Y_{MSM'}/Z_{MSM'})</li> <li>bswx_pred: x component of the FNN prediction for the upstream IMF</li> <li>bswy_pred: y component of the FNN prediction for the upstream IMF</li> <li>bswz_pred: z component of the FNN prediction for the upstream IMF</li> <li>magsw_pred: amplitude of the FNN prediction for the upstream IMF</li> <li>bswx_err: uncertainty in the x component of the FNN prediction for the upstream IMF</li> <li>bswy_err: uncertainty in the y component of the FNN prediction for the upstream IMF</li> <li>bswz_err: uncertainty in the z component of the FNN prediction for the upstream IMF</li> <li>magsw_err: uncertainty in the amplitude of the FNN prediction for the upstream IMF</li> </ol>

opencc-by-4.0Aug 2023View details →

QH9: A Quantum Hamiltonian Prediction Benchmark for QM9 Molecules

This is the official QH9 datasets from paper 'QH9: A Quantum Hamiltonian Prediction Benchmark for QM9 Molecules'.  QH9 is a new  Quantum Hamiltonian dataset providing precise Hamiltonian matrices for 130,831 stable molecular geometries, based on the QM9 dataset. Here is the QH9Stable dataset which is used in QH-Stable-iid and QH-Stable-ood.

Datasets for "Machine-Learning-Enhanced Symbolic Regression for Methane Storage Prediction in Covalent Organic Frameworks"

This collection contains the datasets and associated files used in the research presented in the manuscript titled "Machine Learning-Enhanced Symbolic Regression for Methane Storage Prediction in Covalent Organic Frameworks". The datasets are critical for the development and validation of machine learning and symbolic regression models aiming to predict methane storage capacities in covalent organic frameworks (COFs). Included Datasets: <ol> <li><code>COF_Data_for_ML.csv</code>: This dataset was utilized for the development of machine learning models.</li> <li><code>COF_Data_for_SISSO.csv</code>: This dataset was employed for the development of SISSO-based symbolic regression models.</li> <li><code>ML_vs_GCMC.xlsx</code>: This comparative dataset features GCMC-calculated results alongside machine learning predictions.</li> <li><code>Feature_Combination.xlsx</code>: This file contains data detailing all the feature combinations explored in the study.</li> <li><code>ML_SISSO_GCMC.xlsx</code>: This comparative dataset includes GCMC calculations, SISSO-based symbolic regression model predictions, and ML predictions.</li> <li><code>Crystallographic_Properties_of_535k_COFs.xlsx</code>: This consolidated dataset presents the crystallographic properties of 535,293 COFs.</li> </ol> Software Used: <ul> <li>Machine Learning Computations: Scikit-Learn (<a href="https://scikit-learn.org/stable/" target="_new">https://scikit-learn.org/stable/</a>)</li> <li>GCMC Simulations: RASPA2 (<a href="https://github.com/iRASPA/RASPA2" target="_new">https://github.com/iRASPA/RASPA2</a>)</li> <li>SISSO Calculations: SISSO toolkit (<a href="https://github.com/rouyang2017/SISSO" target="_new">https://github.com/rouyang2017/SISSO</a>)</li> <li>Crystallographic property calculations: Zeo++ (<a href="https://www.zeoplusplus.org/" target="_new">https://www.zeoplusplus.org/</a>)</li> </ul> The datasets are provided to enable replication of the study's findings, encourage further research in the field, and facilitate the development of advanced predictive models by the scientific community. Researchers who use these datasets are requested to cite this Zenodo entry as well as the associated paper upon its publication.

opencc-by-4.0Apr 2024View details →

Predicted Transformation Products of 321 Pesticides

This dataset contains transformation products predicted for 321 pesticides detected in European and Australian surface and ground waters. All predictions were done using the environmental microbial degradation module of BioTransformer3.0 version 2023-05-23. Five generations of transformation products were predicted. The files in this deposition include: pesticides_compounds.sql contains the all unique compounds and their identifiers (including both precursors and transformation products). pesticides_dead_end_tps.sql contains the IDs of all transformation products for which no further transformation products could be predicted as well as the original precursors. pesticides_input_pesticides.sql contains the IDs of the 321 original pesticides. pesticides_reaction_types.sql contains the unique enzyme biosystem and transformation rules. pesticides_reactions.sql contains the unique combinations of precursor IDs and transformation product IDs. pesticides_routines.sql contains query functions to obtain TPs and precursors as well as full transformation pathways. pesticides.sql is a self-contained file containing all the other files.

opencc-zeroApr 2024View details →

Accounting for nonlinear responses to traits improves range shift predictions

Accurately predicting species' range shifts in response to environmental change is paramount for understanding ecological processes and global change. In synthetic analyses, traits emerge as significant but weak predictors of species' range shifts across recent climate change. These studies assume linear responses to traits, while detailed empirical work often reveals trait responses that are unimodal and contain thresholds or other nonlinearities. We hypothesize that the use of linear modeling approaches fails to capture these nonlinearities and therefore may be under-powering traits to predict range shifts. We evaluate the predictive performance of approaches that can capture nonlinear relationships (ridge-regularized linear regression, support vector regression with linear and nonlinear kernels, and random forests). We apply our models using six multi-decadal range shift datasets for plants, moths, marine fish, birds, and small mammals. We show that nonlinear approaches can perform better than least-squares linear modeling in reproducing historical range shifts. Consistent with expectations, we identify dispersal and climatic niche traits as primary determinants of distribution shifts. Traits identified as important predictors and the direction of trait effects are generally consistent across models but there are notable exceptions. Among important predictors, there are more consistent responses to climatic niches than dispersal ability. Modest improvements in predictability when accounting for nonlinearities and interactions and the overall low amount of variance accounted for by trait predictors suggest limits to trait-based statistical predictive frameworks.

opencc-by-4.0Apr 2024View details →

Using neural networks to predict hurricane storm surge and to assess the sensitivity of surge to storm characteristics.

Data for "Lockwood, J. W., Lin, N., Oppenheimer, M., & Lai, C.-Y. (2022). Using neural networks to predict hurricane storm surge and to assess the sensitivity of surge to storm characteristics. Journal of Geophysical Research: Atmospheres, 127, e2022JD037617. https://doi.org/10.1029/2022JD037617"