Benefits of using RO-Crates to publish Galaxy analysis results

Benefits of using RO-Crates to publish Galaxy analysis results Abraham Nieva De la Hidalga (Catalysis Hub, School of Chemistry Cardiff University) Goals Example using Galaxy RO-Crates to support reproducibility Distributed reproducibility environment Describe how we want to extend Galaxy RO-Crates to include domain metadata workflow dataset produce metadata contains indexes indexes invokes references Reproducibility Spectrum Ran once Constraints: -Researcher -Software -Data Repeatable Not repeatable Constrains: -Researcher -Data -Equipment -Software Runnable Constraints -Software -Data Reproducible Supports further development -New software -New Data Replicable Publishing Data Time and effort Extendable New Ideas Same data New data Same researcher New researcher 10.1016/j.envsoft.2020.104753 and https://jhudatascience.org/Reproducibility_in_Cancer_Informatics 3 - + XAS analyses as an example •XAS analyses are important tools for catalysis research •New developments will produce larger quantities of data •Increase the time and resources for processing and analysis Brighter Faster More coherent Higher throughput More data to process and analyse Challenges in Ensuring Reproducibility of XAS results Challenges in Ensuring Reproducibility of XAS results •naming conventions, •use of spreadsheets •different versions of software •data formatting for different tools Need for New Tools for XAS Goal: •Accelerate processing while enabling reproducibility and reusability of results Problem: •Interactive tools do not facilitate processing large datasets and preservation of data requires manual curation. Proposal: •Create/adopt tools which automate part or the whole workflow and produce curated datasets GALAXY Workflows •Web Based •No setup required •Suitable for novice and experienced users •No programming or scripting required •Metadata tracked during processing and analysis Scientific Computing Department Materials Galaxy https://materialsgalaxy.stfc.ac.uk/ XAS Galaxy Tools Openly Available STFC Materials Galaxy Galaxy Europe https://usegalaxy.eu/ https://materialsgalaxy.stfc.ac.uk/ 10 RO-Crates Available on Zenodo https://zenodo.org/communities/rexas/records Tutorials to start using Galaxy RO-Crates https://xerte.cardiff.ac.uk/play_22679 Workflows in EuroScienceGateway collection https://workflowhub.eu/collections/13 RO-Crate Use Case workflow dataset produce metadata contains indexes references Processing platform publishing framework Workflow catalogue data catalogue XAS processing tools Distributed reproducibility environment facilitate adoption FAIRness of XAS Galaxy RO-Crates •Findability: published into an open repository, linked to publication, learning materials, processing environments and other catalogues. •Accessibility: licensed, known[open] formats, standards [metadata and publishing] •Reusability: all data in RO-Crate is accessible from the Zip file or from Galaxy on Restore •Interoperability: Can be executed in different Galaxy environments, if they have the XAS Galaxy tools installed. Can we improve this? •Findable: add domain metadata to facilitate discovery •Accessible: allow open standard formats in all I/O files •Reusable: Still require familiarity and time for identifying data files and the processing hierarchy •Interoperable: Data can be used in other tools (Athena, Larch), bust still requires manual selection. LEVEL UP Why Extend Galaxy RO-Crates? We would like to test different scenarios: 1. Reproduce the results, outside of Galaxy: enable using AiiDA, PWD, Nextflow, Jupyter, Larch, Demeter, or any combination of them for reproducing the experiment. 2. Reuse the data: ensure all data objects are accessible, identifiable and reusable. 3. Reuse the workflow: extract operations, sequencing and parameters to automatically build a “workflow twin” which can execute the operations of the original workflow and provide equivalent outputs. 4. Data driven processing: read the RO Crate, determine the types of data objects contained and which operations can be performed on those outputs (plot, categorise, further processing). Domain Metadata RO-Crate Extensibility https://www.youtube.com/watch?v=7XSQKJOpZ20 Eli Chadwick Common XAS processing and analysis Workflows Detailed view of the XANES Workflow Ingesting and producing CDIF-4-XAS annotated data CDIF-4-XAS manager: handle nxxas (and nxxasproc) and XDI data as inputs and outputs. Inputs: Raw data from beamlines nxxas and XDI. The minimum input should include energy and transmission data. Outputs: nxxasproc or XDI. At minimum it should add a normalized data column as output. Provenance (context metadata): •Experiment details (beamline, facility, date, source name and id) •sample details (element, element edge, edge energy). Integrate additional metadata •On Postprocessing: •Create a list of XDI or Nexus Data and Metadata pairs in the RO Crate •Use workflow provenance data to order in source-produce list •Get JSON-LD context metadata and add it to RO-Crate definition (add to ROCrate zip file) •Generate external file with JSON-LD metadata for publishing •Publishing: •Publish RO-Crate with discovery profile externalised (JSON-LD file) •Test on PSDI invenio •Interoperability •Design process to read extended RO-Crate to use in alternative platforms (Larch, Demeter, Nextflow, Jupyter) RO-Crates •Backed by more than 10 years of development •Built on open, technology-agnostic standards •Metadata generated using existing ontologies •FAIR-compliant data objects with minimal investment •Adaptable to specific domains via RO-Crate profile RO-Crates provide a practical framework for connecting data, standards, and domain knowledge. Alternatives to RO-Crates •BagIt •Focus: Packaging and transfer of digital content •Widely used in digital preservation •Frictionless Data Packages •Focus: Tabular data and lightweight metadata •Strong in open data publishing •OAI-ORE (Object Reuse and Exchange) •Focus: Aggregation of web resources •Based on RDF and linked data principles •DataCite Metadata Schema •standardized description for citation and discovery. •Used on repositories and DOIs. •Focused on interoperability. Facilitating Reproducibility in Catalysis Research with Managed Workflows and RO-Crates: A Galaxy Case Study P. AUSTIN STFC - Scientific Computing Department L.LIBORIO S. DEVASAM A. BELOZEROV T. UNDERWOOD UK Catalysis Hub C.R.A. CATLOWA. NIEVA M. WILDING Diamond Light Source N. RAMANAN Galaxy Workflow Management System Service CDIF-4-XAS Project Who is doing it? -CODATA, the Committee on Data of the International Science Council, France. Simon Hodson (PI), Arofan Gregory, Matti Heikkurinen -Rutherford Appleton Laboratory, STFC, UKRI, UK. Leandro Liborio, Patrick Austin -School of Computer Science and Informatics, Cardiff University, Wales, UK. Abraham Nieva de la Hidalga -Helmholtz-Zentrum Berlin für Materialien und Energie GmbH, HZB, Germany. Heike Görzig, Rolf Krahl -Helmholtz Metadata Collaboration (HMC) Hub Matter, Germany. Markus Kubin Do RO-Crates fit in your data ecosystem? •Do RO-Crates complement or conflict with existing standards? (BagIt, Frictionless, OAI-ORE, or DataCite) •Challenges of integrating semantic metadata into data packaging •Are RO-Crate profiles sufficient for the chemistry domain? Is further customization needed? •Role of ontologies evolving in practical data sharing Reproducibility in Chemistry