Let's Talk About It: To Disclose or Not to Disclose?
TRANSMART NOV 2013 MEETING REPORT
1. TRANSMART NOV 2013
CHILLY-MAZARIN, FRANCE
Yi-Ke Guo, Imperial College – y.guo@imperial.ac.uk
Kees van Bochove, The Hyve – kees@thehyve.nl
2. Since June 19, 2013 …
tranSMART 1.1 has been released (15th/Sept)
> 100 man/month effort
~4oo code commits have been made into
tranSMART repositories
>30 wiki pages created or updated
>30 discussions started on the mailing group
15 developer meetings have taken held
3 community webinars taken place
7 member organisations joined tranSMART
Foundation
… and many more
3. Where are we in tranSMART?
Joint efforts by Pharma, Academic, for-profit
service providers and software venders
Fully open source based
Rigorously tested ( > 500 bugs fixed)
With ETL scripts
Documentation
Public hosting
4. tranSMART V1.1 as the Base
tranSMART V1.1 is extended to support both
Oracle and Postgres
Major efforts has been made towards a
modular architecture with APIs
tranSMART V1.2 will be based on these
efforts to integrate various separated
developed “local versions” into one common
platform
5. Lots of Things Learned
How to coordinate
How to balance development and release
How to promote a process within the
community
How to keep innovation while delivering the
needed production system
How to develop a community
How to make everyone “almost” happy
6. Lots of Challenges Ahead
Product management: requirements from a board
user space with different priorities and
expectations
Code management: the shared branch for release
and the local development branches
Process management: centrally coordinated
release process and distributed development
process
8. Where do we want to go?
- an open question for this meeting
- some ideas:
Progress from ‘showing to each other what
we do with tranSMART’ to really work
together
Identify the overlap in the multiple projects
and stakeholders: what brings us really
together, as scientists, as developers?
Convert ideas into action by setting ourselves
concrete goals with small dedicated groups
9. #1 People do great things
with tranSMART
The fact that we are all
here (again) shows we are
rallying around a common
cause!
10. #2 TranSMART brings together
parties that would otherwise
not work together on a dayto-day basis
Amsterdam:
>10 pharma companies
>10 universities/medical centers
>10 IT companies
Paris:
Academic/Research Institutes : 18
Pharmaceutical Companies: 10
IT/Service companies : 19
11. #3 TranSMART needs to grow
as a product.
Opinion: To do that, we need
to identify what is common
(‘core’), and let the rest
breathe!
12. #4 The TranSMART developer
community needs to grow a
joint working model around
the core.
This has very demonstratably
happened in the last months,
around the 1.1 release!
13. #5 TranSMART has great
potential as a data sharing
platform.
But we need to make it
easier to get data in first.
14. #6 The TranSMART community
needs to develop a working
model for Quality Assurance.
At least from a development
standpoint.
15. #7 TranSMART needs a plugin
repository, to survive as an
open source project.
It’s starting – come to the
hackathon!
16. #8 TranSMART has two
distinct user communities:
clinicians and
bioinformaticians.
We might need 2 user
interfaces!
17. #9 TranSMART unites clinical
data from EHR and eCRF
systems, and bioinformatics
data from many sources.
But it has 2 distinct study
types: longitudinal studies
vs clinical trials /
experiments.
18. #10 TranSMART would benefit
from a Community Manager –
someone whose fulltime job
it is to connect people and
ideas in the community.