Bioschema metadata discussion in the biohackathon

[NishaSharma14]

nmrXiv team was a part of 1st BioHackathon Germany in the project of (Bio)Schemas4NFDI, lightweight domain metadata (not only) for NFDI consortia. Here you can find all the points that were solved where your comments and further improvements are welcome, along with the points that are still open for discussion.

• Solved points:

  • In an nmrXiv assay, should we add the solvent details to the dataset of the sample?
    • This should depend on the source of information. If the solvent was extracted from the NMR files that are dataset specific, it should be added to the dataset. In case it comes from the protocol of sample preparation, it should go to the sample.
  • Sample will be deprecated in favor of [BioSample] (https://bioschemas.org/types/BioSample/0.1-RELEASE-2019_06_19) (BioSample is already in nmrXiv)
  • How to use study schema for projects?
    • This can be solved by using "conforms to" and mappingto ISA investigation.
  • In nmrXiv we have Study about sample, while MassBank has dataset about molecularEntity. We believe that both usages are fine due to the difference in the structure between nmrXiv and MassBank (no samples in MassBank)
  • Repositories in NFDI4Chem will go with the MolecularEntity and will avoid ChemicalSubstance as the first covers more molecular details and identifiers

• Open points:

  • How best to add the conditions/results of the assay to the dataset?
    • MassBank adds those details to "variableMeasured".
    • This solution is not ideal, as the "variableMeasured" should refer to the target measured in the assay rather than its conditions. However, at the moment this is the best way to do it as it can have ontology terms, which is useful for the search, even if they don't fit in the right property. Bioschema organizer recommended this approach too. For the future, we aim to add more properties to the dataset, including the conditions and results.

• Validation

  • schema.org validator
  • BioSchema validator Currently down
  • FAIR Checker

To Join the main channels for Bioschemas, and their community.

1. Subscribe to the mailing list
2. Get involved in discussions on GitHub
3. Join the slack space
4. There is also a community call once a month and reminders will be sent to the mailing lists

• Tasks:

  • LiveDeploy to BioSchema
  • Get in contact with dataset group on github to inlude conditions and results
    once we have the supported & validated schema, we may want to add it to live deploy page.
    https://github.com/BioSchemas/bioschemas.github.io/blob/124c4f19b272ba46859bf3c499065a1c0821630b/_data/live_deployments.json
    by submitting a PR to the config file.
    https://github.com/BioSchemas/bioschemas.github.io/blob/124c4f19b272ba46859bf3c499065a1c0821630b/_data/live_deployments.json
  • BioSchema Markup
    Rene implemented this in MassBank/MassBank-web@44790c6
    using the Java GSON library for proper encoding. For PHP there could be the equivalent https://github.com/tebru/gson-php
  • DataCatalog:
    • Add version
  • Project
    • conform to ISA
  • Study
    • conform to ISA
    • Re-define the structure the output json. May be the study details can be displayed first followed by project and dataset details. May be isPartOf is something we can use.
  • Dataset
  • use "variableMeasured" for NMRium info and results
  • Add "distribution" when available

• Future tasks:

  • Add "distribution" when available
  • Add version when available

[NishaSharma14]

To look for json schemas:

1. Go to https://discovery.biothings.io/registry
2. Go to the tab "Browse by Namespace"
3. If you are looking for a release you can go to "Bioschemas" (https://discovery.biothings.io/view/bioschemas)
4. If you are looking for a draft you ca go to "BioschemasDrafts" (https://discovery.biothings.io/view/bioschemasdrafts)
5. Find the profile you are interested in, for instance Dataset (released version at https://discovery.biothings.io/view/bioschemas/bioschemas:Dataset) and navigate to the page
6. On the top right, there is a link "View Source"

For markup examples

1. Go to https://bioschemas.org/
2. Navigate to the profiles list from the menu Specifications-Profiles (https://bioschemas.org/profiles/)
3. Stay on the tab "Release" or go to the tab "Draft" depending on the profile you are interested in
4. For instance, if you are interested in Dataset, you would end up at https://bioschemas.org/profiles/Dataset
5. On the top, go to the tab "Links" (https://bioschemas.org/profiles/Dataset/1.0-RELEASE#nav-links)
6. Look for the "Examples" link and there you will find some examples, commonly corresponding to the minimum properties

Alternative

1. Go to Live Deploys and navigate by Profile (https://bioschemas.org/developer/liveDeploys#nav-profile)
2. Find the profile you are interested in
3. Open the page for any of the listed resources
4. Use Developer Tools in your browser and locate the piece in the HTML corresponding to a script "application/ld+json"
5. You should get the JSON-LD markup which you can copy to an editor to have a better look

[NRayya]

Most of the decided features are already provided. Others will be discussed more in the TA15 regular meetings