Overview

Teaching: 15 min
Exercises: 0 min

Questions

Why submit my data to a repository?

What types of repositories are there?

How do I find a suitable repository?

Objectives

Explain why data should be publicly available.

Explain different types of repositories and how to find a suitable one.

Why submit your datasets to a repository?

Open Science & FAIR - To meet the requirements from funders and society on Open Science & FAIR
Reproducibility - So that your published research results can be reproduced
Trail of evidence - To provide a provenance of the data
3rd party access - To give others access to your data
Archival purposes - Research data should be available for as long as it is useful to someone
Publication of paper requires it - Nowadays most publishers require you to submit the data to a repository when publishing a paper

FAIR data

Data publication is the best way to make your research projects FAIR since your data becomes:

Findable by being assigned a persistent identifier, and by being described with rich metadata.
Accessible by being put in a resourse that is searchable, and enables easy access via internet
Interoperable by using standard format and language to represent both the data and its metadata
Reusable by fulfilling the F, A, and I, and by having a clear and accessible data usage license

Repositories provides the technical solution to FAIR data. Hence, by submitting data to a repository, your data becomes FAIR and you don’t have to provide a solution on your own.

Note that, while we focus on life science research data, the same principles apply to any other type of research data.

What data should be submitted?

Raw data: this is the data that comes straight from the instrument, e.g. RNA sequences in fastq format
Processed & analysis data: this is the data where some type of analysis or processing has been done, e.g. normalization, removal of outliers, expression measurements, statistics
Metadata: this is the description of the raw and processed data, e.g. in the form of minimum information to reproduce the data, sample information, precise protocols

How to find a suitable repository

Types of repositories

Domain-specific:
- Best choice if suitable, long-term plan, typically free of charge, maximum reach
- E.g. European Nucleotide Archive, ArrayExpress, PRIDE
General purpose:
- Second best, long-term plan, might cost (now or in future), good reach but less specific in metadata → more difficult for future users to judge if a dataset will be useful
- E.g. Zenodo, (SciLifeLab) Figshare, Dryad
In-house/institutional
- For archive/backup purpose mainly, might cost, limited reach unless also published in data catalogue

How find a domain-specific repository?

EBI repository wizard - guide depending on data type
ELIXIR deposition databases - core resources with long-term data preservation and accessibility plans
FAIRsharing.org/databases - catalogue of many repositories, with possibility to filter on e.g. domain
Scientific Data Repository Guidance - publisher’s recommendation

EBI Repository Wizard

EBI hosts several life science repositories, suitable for different types of data. The Repository Wizard helps to identify which one is suitable for your data.

Go to the Wizard at https://www.ebi.ac.uk/submission/

Picture
Either explore the wizard with the purpose of finding a suitable repository for one of your projects, or choose among the scenarios provided below. Which repository is recommended?
- Genomics project with RNA sequences
- X-ray crystollography structure of a protein
- Gene expression data
- Protein sequencing data
- Proteomics project using mass spectrometry
- Electron microscopy structure images
Solution
- Genomics project with RNA sequences: European Nucleotide Archive (DNA/RNA sequence -> no controlled access -> produced experimentally -> Other)
- X-ray crystollography structure of a protein: wwPDB OneDep (Structures -> X-ray crystollography)
- Microarray gene expression data: ArrayExpress (Expression data -> no controlled access -> Microarray gene expression)
- Protein sequencing data: UniProt SPIN (Protein data -> no controlled access -> produced experimentally -> Protein sequencing)
- Proteomics project using mass spectrometry: PRIDE (Protein data -> no controlled access -> produced experimentally -> Mass spectrometry -> Proteomics)
- Electron microscopy structure images: EMPIAR (Structures-> Electron microscopy -> micrographs or particle stacks)