Share your research data

The data presented here support the research paper “Provenance Modelling of Fossil Dinosaur Bones Using Geochemistry and Machine Learning”, intended for a submission to "Paleoworld". The dataset contains trace elements concentrations from fossil dinosaur bones from the Upper Cretaceous Nemegt and Djadokhta formations. For the analytical purposes, the dataset was divided into two subsets: the first one consisting of long bones (tibiae, femora, radii and humeri) and the other including trabecular bones (ribs and vertebrae) and metatarsals. Locality labels were used to train and evaluate several machine learning classifiers (logistic regression, random forest, AdaBoost, XGBoost) to assess the potential of bone geochemistry for provenance prediction. Feature selection was conducted on the best-performing models to identify the elements contributing the most to the model performance. These results were compared with those obtained using Linear Discriminant Analysis. The data are provided in CSV format in the “Data” folder. The folder “Plots and figures” contains the figures used in the manuscript, including the plots. The folder “Supplementary files” contains additional files. These files are: - interactive HTML plot ("Element profiles.html") showing the all the concentration profiles across each analysed sample, including the ones measured along several profiles - concentration profiles presented in a PDF file ("All profiles.pdf") - XLSX file with the statistical summary of the data ("Data description WK.xlsx") - LDA scalings in CSV file ("LDA scalings.csv") - The tables comparing predictions and performance of the algorithms using test part of each subset ("Predictions and accuracy.xlsx") Besides that, the Jupyter Notebook with data analysis is also provided ("Bones from Gobi - loc prediction.ipynb").

original date

Emerging evidence suggests that amino acid homorepeats (HRs) in proteins (HRPs) contribute to protein interactability. What is the role of HRs in human-pathogen protein interactions? We find that pathogens engage physiologically important human HRPs, thereby affecting diverse host physiological processes. From the pathogen standpoint, (i) eukaryotic pathogens engage more HRPs but with host-sparse HRs, leading to disparate and discriminate interactions, (ii) prokaryotic pathogens engage less HRPs but with host-abundant non-polar HRs via host protein proxies bringing about discriminate or promiscuous interactions and (iii) viral pathogens engage more HRPs with host-abundant polar uncharged HRs affecting promiscuous interactions using host-partner HR tract mimicry. To propel further research, we introduce a resource Hi-PHI (http://hiphi.iisertirupati.ac.in/) cataloging critical information about human and pathogen HRPs and HRs. We propose mechanisms to (i) repurpose drugs targeting human HRPs engaged by pathogens for treating different infections and (ii) exploit HRs and their flanks as targets for pathogen-targeted anti-infectives. Here, we have uploaded the assembled and curated human-pathogen protein interactome (HPI), which has 19,535 interactions between human and pathogen proteins. We have also provided the source code to facilitate repetition of this work and address other fundamental systems- and molecular-level questions. The instructions regarding usage of the codes are provided in individual scripts. All the datasets assembled, curated, generated and used in this study is available as a resource, Hi-PHI database (http://hiphi.iisertirupati.ac.in/).

The data of ferromanganese nodules.

Folder "High symmetry" contains raw data for carbon nanobracelets made of 2-5 monomer fragments. Folder "Low symmetry" contains raw data for carbon nanobracelets made of 4-5 monomer fragments. ├ ARC files - result of PM3 optimization by MOPAC2016 with symmetry constraints for high-symmetry molecules and without symmetry constraints for low symmetry molecules, ├ OUT files - result of DFT optimization by PRIRODA which uses for input configuration the PM3 pre-optimized geometry; should be opened by Chemcraft software to see list of optimization steps and frequencies for optimized geometry, ├ PNG files - visualization of DFT optimized geometry by Chemcraft software, └ XLS files - calculation of the molecule outer radius, the variation of bond lengths in chains by Eq. (1), and interchain distances shown in Figure 2. Folder "MEP Jmol" contains Jmol visualizations of the Molecular electrostatic potential (MEP) of the studied molecules. ├ XYZ files - input files containing PRIRODA optimized geometry (three columns XYZ; in angstroms) and partial atomic charges (last column; in elementary charge e), ├ SPT file - Jmol script used for MEP visualization; scale factor e²/4πε₀Å = 1.439964548·10⁴ meV converts e/Å (Jmol MEP units) to meV, ├ PNG files - MEP visualizations, ├ JAR file - Jmol v. 14.31.41; 2021-05-29, └ BAT file - command/batch file for execution Jmol under Windows environment. Folder "Molden" contains Chemcraft visualizations of HOMO and LUMO of the studied molecules. ├ OUT files - PRIRODA single point Molden calculations of the optimized molecules containing only HOMO and LUMO orbitals prepared by the GAWK script from the large (~100-500 Mb) output, ├ PNG files - HOMO and LUMO visualizations; render settings, ├ TXT files - Camera settings for HOMO and LUMO visualizations, ├ AWK file - GAWK script used for preparing short OUT files containing only HOMO and LUMO orbitals, ├ ZIP file - GAWK v. 3.1.0-2; 2001-06-03, └ BAT file - command/batch file for GAWK script execution under Windows environment.

Fatigue-Constrained Optimisation of Offshore Wind Jacket Substructures Using the HPEO Algorithm – Input Data and Supplementary Material Description: This dataset contains all numerical inputs and supplementary materials associated with the article “Fatigue-Constrained Optimisation of Offshore Wind Jacket Substructures Using the Heuristic Particle Elimination Optimisation (HPEO) Algorithm.” The data support the findings reported in Section 6 of the manuscript and enable full reproduction of the optimisation studies. Purpose and Reuse Notes The dataset is intended to facilitate replication of the fatigue-constrained jacket optimisation study, enable benchmarking of future discrete optimisation algorithms, and support further exploration of joint-level hot-spot stress modelling in offshore wind applications. All files are provided in open, human-readable formats to maximise reusability. Researchers may reuse or extend the dataset under the associated licensing terms, with appropriate citation of the original article.

This dataset contains the custom real-world image dataset and modified source code used to benchmark Neural Radiance Fields (NeRF) with stratified sampling against the original random sampling method. It supports the reproducibility of the findings in the associated manuscript.

Knapsack Problem instances for N = 50 generated with various Quality Diversity Methods. Each instance includes its target solver and the method used to generate it.

Unprocessed Western Blots and Counting Data

The data serves a comparison of the valuation of urban environmental resources with stated preference methods, i.e. willingness to pay (WTP), and willingness to accept (WTA), applied the Kaskada Park in Warszawa, Poland and the Decjusz Park in Kraków, Poland. The results obtained indicate that respondents value various aspects of urban parks; however, assigning a monetary value to these aspects is not always straightforward and often poses a challenge. The survey shows that the valuation of environmental resources based on the stated preference methods generally confirms that these resources are considered valuable. The diagnostic survey, based on a questionnaire tool, was hosted on an online platform. It was distributed through social networks and disseminated via email as well as through QR codes at local institutions and businesses. The same survey was used for both parks in order to identify differences and similarities in willingness to pay and willingness to accept compensation among residents of Warszawa and Kraków.