We have a few use cases that I would like to run by someone on Synapse's end before proceeding. We don't expect them to be problematic, but want to confirm beforehand.
Is there a point of contact or team that I can reach out to?
Thanks in advance,
Chris
Created by Chris Rhodes crhodes4 Hello @crhodes4,
Thanks for the context and we appreciate your caution about sharing this data! I don't think that I'm the best person to answer these questions, but I will point you to our [Access & Compliance Team (ACT)](https://sagebionetworks.jira.com/servicedesk/customer/portal/8), where you can use the "Other questions" option. Since this help desk is geared towards concerns or questions about data access and sharing, I think that this should get your questions to the right people. Please circle back if this is not the case and you need me to reach out to other teams at Sage.
Best,
Will
HI Will
@wpoehlm,
A team member passed along what I think is a great example of our desired primary use-case:
Olah M, Menon V, Habib N, Taga MF, et al. Nat Comms, 2020. https://www.nature.com/articles/s41467-020-19737-2.
The authors use Synapse hosted human microglia data (syn21438358) to populate browsers:
https://vmenon.shinyapps.io/microglia
https://cellxgene.cziscience.com/collections/fcb3d1c1-03d2-41ac-8229-458e072b7a1c
I expect the reason such applications are be acceptable is that the end-users do not have access or knowledge of the underlying controlled counts matrices and other sensitive info?
To be clear, we DO NOT plan to make any apps/portals available to the general public, unless directed by Synapse as supplementary info for publications.
Best,
Chris HI Will @wpoehlm,
Our primary use-case that we need advice on:
We have a team of certified Synapse users. Our synapse users would like to combine controlled-access single cell RNA-Seq data (ex syn30961086) with publicly available, open access data (ex cell-by-gene matrices from NCBI GEO) to generate larger disease atlases. Once these atlases (derived in part from controlled-access files) are created, we would like to make the data available to our larger organization in the form of a private app, such as through ShinyCell (https://github.com/SGDDNB/ShinyCell) for visualization, exploratory stats, etc.
The main concern we have is whether we can share such data within our organization, which is primarily non-Synapse users. It is not realistic to add potential users to our current synapse DUC on a rolling basis. How would you suggest doing this?
As an alternative to the use-case above, if we created such controlled/open access atlases and then trained a SCVI model, or equivalent, I expect sharing the model (but not underlying controlled counts matrices) would not violate terms of service, correct? This is indeed one of the goals of such ?transfer learning? models. This would limit the use of such data internally but would at least have some utility.
More specifically, it seems we restricted from sharing ?locally stored copies? of files (e.g those local files that have md5sums matching corresponding remote files on Synapse.org), correct? So if we derive files, such as Differentially Expressed Gene lists from controlled access data, those are not considered controlled access?
Let me know if I need to clarify anything.
Thanks,
Chris
hello @crhodes4,
Can you provide more context and details about these use-cases? I might be able to address them, or I can direct them to the right people here.
Thanks,
Will