Submission of datasets to the IDR

We aim to publish reference image datasets, which have value beyond simply supporting an original publication according to the Euro-BioImaging - Elixir Image Data Strategy. This includes:

• Datasets associated with an existing or upcoming publication
• Complete datasets - not just images supporting one figure in the publication
• Datasets whose metadata can be integrated with other datasets via identifiers from well-known biomolecular resources (Ensembl, NCBI Entrez Gene, RefSeq, PubChem, ChEBI etc)
• Datasets generated using new imaging methods or new analysis methods
• Datasets that are likely to be re-analysed or incorporated into other studies or integrated with other imaging datasets

To make IDR datasets as widely re-usable as possible, we strongly recommend that submitters make their datasets available under CC-BY license. The licensing information should be included as part of the metadata .

Dataset size is typically not an issue, but for sizes significantly larger than 1000 GB special planning may be needed.

Metadata describing an imaging study is submitted using template files. These are available for download from https://github.com/IDR/idr0000-lastname-example/archive/master.zip.

There are template files for two different use cases:

1. Templates for high content screens (HCS), or imaging studies performed in plates (e.g. 96 or 384 well plates)
2. Templates for non-screen datasets, or studies which group images into a number of distinct experiments

There are help pages to provide more information about filling out the high content screen and non-screen dataset templates.

For both types of study three files should be submitted:

• Study file - describing the overall study and the screens or experiments that were performed
• Library (screens) or assay files (non-screens). For HCS the library file describes the plate layout of each screen. If more than one screen is performed as part of the study there should be a library file for each screen. For non-screen datasets, the assay file(s) describes each imaging experiment included in the study.
• Processed data files - containing summary results and/or a ‘hit' list for each screen or experiment. It may include phenotypes observed.

Example files for HCS studies can be found here:

• idr0001 - gene deletion screen
• idr0002 - RNAi screen
• idr0006 - protein localization screen
• idr0009 - primary and validation RNAi screens
• idr0017 - small molecule screen
• idr0019 - imaging experiment performed in plates but not a screen

Example files for non-screen studies can be found here:

• idr0018 - histopathology of gene knockouts
• idr0021 - protein localization using 3D-SIM
• idr0023 - protein localization using dSTORM
• idr0027 - 3D-tracking of tagged chromatin loci
• idr0032 - in situ hybridzation of RNA

Each study is slightly different and advice on how to fill in the template can be given. Questions should be sent to idr@openmicroscopy.org.

Imaging datasets can be remarkably complex. To the extent that it's well-structured and reusable by the community, we are interested in enriching your online dataset with your additional metadata, e.g.:

• Rendering settings: if you have information on the channel intensities or mappings that you would like to see displayed in the IDR, please let us know.
• Regions of interest: if you have ROIs or masks, whether detected automatically or hand-annotated, we'd like to make them available in the IDR.
• Features: if you've calculated features for your dataset for machine learning, we'd be interested in sharing those analysis results with others.
• Pipelines and workflows: further, if you have a representation of how you ran the analysis itself, we think that should be included with your features.

There are several options for having your image data transferred to the IDR:

• Provide a link: if your data is already available online (e.g. Dropbox, FTP, etc), a link to the location may be the easiest option.
• FTP: data can be uploaded using our FTP server. We will provide you with details via email.
• Shipping disks: Sending hard-drives by mail remains a reliable way of transferring your data. We will provide you with shipping details via email.
• FTP / Aspera: we are working on a link between IDR and BioStudies ( https://www.ebi.ac.uk/biostudies/submit.html). If you can upload your image data to BioStudies via FTP or Aspera, IDR may be able to access your data directly from there.

A submission is considered complete only after your uploaded metadata and image data have been validated. Validation may itself take significant time depending on, e.g., the size of the dataset, the file formats used and the consistency in naming between files. During validation, frequent consultation with the author and/or data producer may be necessary.

Once your data is validated and ready for import, you will be notified that the publication pipeline can begin.

• Allow at least 3 weeks before your dataset will be available online
• No private access is available for reviewers during this period
• Minting a DOI requires approximately 1 additional week (Note: if you plan on incorporating a data DOI into your publication, please allow time for adding this DOI into the final version of your paper).

Your dataset in IDR will be given an accession number (e.g. “idr1234”) which you can use for informally referring to your entry. For more formal uses such as citing the imaging dataset in your publication, we suggest using a DOI:

Image files are available in the Image Data Resource (Williams et al 2017) under DOI https://doi.org/10.17867/10000105 .

DOIs can be made available upon request but require an additional depositor agreement and license form on top of your completed submission.

If you want to include a citation to the IDR in a paper or other publication, please use:

Williams et al (2017) Image Data Resource: a bioimage data integration and publication platform, https://doi.org/10.1038/nmeth.4326.

Email idr@openmicroscopy.org with some information about your dataset and timeline for publication.