RDM in systems neuroscience:

Challenges and best practices
Adina Wagner
mas.to/@adswa


Institute of Neuroscience and Medicine (INM-7),
Research Center Jülich

Institute for Experimental Psychology,
HHU Düsseldorf
Slides: files.inm7.de/adina/talks/html/fdm_nrw_hhu.html
DOI; 10.5281/zenodo.10122803

Perks of being a neuroscientist...

A growing culture of open data https://en.wikipedia.org/wiki/List_of_neuroscience_databases

Perks of being a neuroscientist...

A large and growing amount of open source software



... and many more!

Perks of being a neuroscientist...

Many readily available, often free, sometimes FOSS, services for data storage and collaboration
https://www.incf.org/resources/sbps

Perks of being a neuroscientist...

Work on fascinating questions with fascinating data
Dwayne Reed; commons.wikimedia.org
/wiki/File:Parasagittal_MRI_of_human_head_in
_patient_with_benign_familial_macrocephaly_prior
_to_brain_injury_(ANIMATED).gif 		https://mne.tools
https://mne.tools http://trackvis.org

The good news, the bad news

Interesting data ... that's subject to GDPR, making storage, analysis, and sharing difficult
Data analysis Reproducibility is threatened by intransparent, multi-stepped analyses¹ & unstable results across software versions². Dataset sizes exceed computational capabilities (e.g., HCP project: ~100TB, 1.2k people).
Data sharing Heterogenous distribution and updating, many scientists lack data management skills
Collaboration Few interoperable workflows across institutes, rather: isolated solutions

¹ Botvinik-Nezer et al., 2020: Variability in the analysis of a single neuroimaging dataset by many teams
² Kiar et al., 2020: Comparing perturbation models for evaluating stability of neuroimaging pipelines

  • Domain-agnostic data management tool (command-line + graphical user interface), built on top of Git & Git-annex
  • 10+ year open source project (100+ contributors), available for all major OS
  • Born from rethinking data:
    • Just like code, data is not static.
    • Just like code, data is subject to collaboration. Stream-lined workflows for sharing and collaborating should be possible, mirroring those in software development.
    • Provenance of data is essential for reproducible, trustworthy, and FAIR science
    • Flexibility and interoperability with existing tools is the key to sustainability and ease of use

  • Domain-agnostic command-line tool (+ graphical user interface), built on top of Git & Git-annex
  • 10+ year open source project (100+ contributors), available for all major OS
  • Major features:
    • Version-controlling arbitrarily large content
    Version control data & software alongside to code!
    Transport mechanisms for sharing, updating & obtaining data
    Consume & collaborate on data (analyses) like software
    (Computationally) reproducible data analysis
    Track and share provenance of all digital objects
    (... and much more)

Tracking & Sharing - Where is the Privacy?

  • Datasets have an optional annex for (large or sensitive) data.
  • Rather than file content, identity (hash) and location information is tracked. Users have fine-grained control over transport and access; encryption is possible.
  • Datasets can separate data access from meta data access

Fine-grained file transport

  • Cloned datasets are lean, because file content is retrieved on demand:
    •       $ datalad clone \
    https://github.com/datalad-datasets/human-connectome-project-openaccess.git
$ cd human-connectome-project-openaccess && du -sh
  5.0M	.
$ datalad get HCP1200/102513/T1w/T1w_acpc_dc.nii.gz
  get(ok): HCP1200/102513/T1w/T1w_acpc_dc.nii.gz (file) [from datalad...]
$ datalad drop HCP1200/102513/T1w/T1w_acpc_dc.nii.gz
  drop(ok)

Big data

Where can I find out more?

{Open,Transparent,Reproducible} science

  • Treat data like software: obtain, version, share, and update data
  • Simplified data management, disk-space aware storage & computing
  • Transparent and reproducible science: link code, data, software, and execution in a human- and machine-readable way
  • Collaborate: Generic workflows, interoperabality with established tools & services

distribits.live

  • First conference on technologies for distributed data management
  • 2 day conference plus single-day Hackathon
  • @ Haus der Universität Düsseldorf

Acknowledgements

Software
  • Joey Hess (git-annex)
  • The DataLad team & contributors


Thanks!

Questions?
Funders
Collaborators