This document discusses Research Objects (RO), which provide a framework for bundling, exchanging, and linking resources related to experiments in order to improve reproducibility. The RO framework uses unique identifiers, aggregation, and metadata to group related resources. Real-world examples of ROs include reviewed scientific papers, workflow runs, and Docker images. ROs can help make research fully FAIR (Findable, Accessible, Interoperable, Reusable). Tools and platforms like FAIRDOM, SEEK, and Figshare support the use of ROs.
3. Framework desiderata
Technology
Independent.
The
least
possible
The
simplest
feasible
Graceful degradation
Standard
tooling
4. How?
The
Container
Packaging:
Zip
files,
Docker
images,
BagIt,
Web,
…
Catalogues
&
Commons:
FAIRDOM
SEEK,
Farr
Commons
CKAN,
myExperiment,
Zenodo,
Figshare,
…
Manifest
Describes the aggregated resources, their
annotations and provenance
Manifest
5. Manifest
Manifest
Construction
• Identification
–
id,
title,
creator,
status….
• Aggregates
–
list
of
ids/links
to
resources
• Annotations
–
list
of
annotations
about
resources
Manifest
Manifest
Description
• Checklists
–
what
should
be
there
• Provenance
–
where
it
came
from
• Versioning
–
its
evolution
• Dependencies
–
what
else
is
needed
Manifest
7. RO Principles
Use unique identifiers as names for things.
Use some mechanism of aggregation to
group things together.
Provide metadata about those things &
how they relate to each other.
11. Reviewed to Reproduced
From González-Beltrán et al. doi:
10.1371/journal.pone.0127612
Reproducibility
Same data
Same code
Systematic and
extensible
meta-data
collection
✔
✔
15. Use
case:
ATLAS
Collider
Data
Analytics
Portable,
lightweight
application
runtime
and
packaging
tool.
Image
ATLAS
and
CMS
detector
data
Charles
Vardeman,
Da
Huo
All
data
and
files
of
the
execution
+
Instructions
convert
bundle
manifest
Relate
files
and
layers
Add
provenance
and
annotations
Link
in
other
content
run
Exchange
Reproducibility
Same data
Same code
Same run time
environment
Systematic and
extensible meta-
data collection
✔
✔
✔
22. Research Objects
Publish a digital record
of your entire scientific
enterprise
You can give it
to someone
else
You can get
credit for it
People think
you are a good
person
You get a
promotion
• Why does this matter to Biologists?
24. Conclusion
• Simple solution, addressing needs towards
transparent FAIR principles
– Findable, Accessible, Interoperable, Reproducible
• Adoption
– Training
• Online tutorials
• Face to face
– Need more tools that take advantage of the RO
Framework and lower the cost (technological
debt) of reproducibility
• Work together
25. Acknowledgements
Carole
Goble
Stian
Soiland-‐Reyes
Matt
Gamble
Rob
Haines
Sean
Bechhofer
Phil
Crouch
Finn
Bacall
Stuart
Owen
Carole
Goble
Khalid
Belhajjame
Graham
Klyne
Jun
Zhao
Daniel
Garijo,
Oscar
Corcho
Esteban
García
Cuesta
University
of
Manchester
University
of
Oxford
Lancaster
University
UPM
http://researchobject.org
http://fair-‐dom.org
http://www.seek4science.org
http://www.farrinstitute.org
http://www.wf4ever-‐project.org
http://myexperiment.org
Raul
Palma
iSOCO
PSNC
Paris
6