Confirmation that data is fully de-identified

Competition Broad Institute Obesity Crunch #1
1

Hi,

I am trying to use our school’s cluster to work on this task, yet they request confirmation from the competition organizers that the data is fully de-identified. May I have a confirmation on that?

Thanks.

2

Hi Chris,

I am not 100% what you mean by that.

The data still have the real gene names.

I suggest you just open the quickstarter (here is another one) and look at the previous cell executions. You can even run it on Colab and see for yourself.

3

Hi,

Sorry for the earlier confusion. To clarify: my school’s cluster admins are strictly concerned with Patient Privacy (PHI/HIPAA).

Could you confirm if this data comes from lab-grown cell lines (like K562, HEK293, etc.) rather than direct clinical samples from human subjects? If you could provide the name of the cell line or a reference to the source study, that would be enough to satisfy our ethics board that the data is de-identified. Thanks

4

Hi Chris,

The official specification document make no reference to such certification.

If you would like, I can contact the organizers directly regarding your question, as I will not be able to answer it myself. However, you may not receive an answer over the weekend.

At the end of the document, there are a lot of references that the organizers recommend reading before jumping into serious work for the competition.

5

Thanks. Since the cluster admins require formal proofs that the data is Non-Human Subjects Research, could you ask the organizers for a quick confirmation of the cell line name or the source study citation?

Again, thanks for your time.

6

I just sent the email. I will forward their response once they reply.