Ingesting the same sample in two different projects

[anjalijain22]

Hello, hope you're doing well!

I'm wondering whether seqr (with the Hail backend) supports ingesting the same sample into multiple projects? For instance, can I upload a VCF to Project A today, and then upload the same VCF to Project B a month later? Would that also lead to duplication of allele counts in seqr for the samples present in both projects A and B?

[hanars]

Hi Anajli,

There is nothing that will prevent you from ingesting the same sample into multiple projects. However, this would lead to duplication of allele counts and thus we strongly discourage doing this, as it adversely affects your searches and the ability to assess variants. Can you explain why this is something you would be considering in the first place? The primary reason for splitting data into different projects is usually access control, but if it is all the same samples then it would seem more logical to keep one project and grant everyone who needs access to it. Understanding the use case may help me propose better alternatives

Best
Hana

[anjalijain22]

Hi Hana,
Those were exactly my concerns as well! The reason we're considering this approach is that we have one family enrolled in two different research studies. We could have a separate project for just that family and grant access to both research studies to that one project, but the family has different IDs in each study. For consistency, it would make the most sense to include them under the main project.

Is there a way to process the genomic data just once with the IDs from Project A, and then link that same data on the backend to Project B? I’m just exploring all possible options, totally okay if this isn’t currently supported.

[hanars]

Linking data from one project to another is not possible in seqr. It seems like your options are to either make a project just for that single family, or you can duplicate the data. There is nothing that will detect or prevent data duplication, its just something you will need to keep in mind going forward as you do your analysis. While it is not recommended, a single family duplicated should not be too terrible for your allele counts. That said, I do think having a project with a single family is also a valid solution, but the researchers will be able to see each others tags and notes.

[anjalijain22]

Hi Hana, that makes total sense, thank you for sharing your insights!

I have a follow up question, I was playing around with the data loading interface and found that I can select a VCF and select multiple project to load it into. For example, here I am able to select two projects entitled "Test: Ashkenazi Trio" and "Test: Ashkenazi Trio-2". Would this approach of loading a sample into two projects also cause allele duplication? If not, I guess the only thing that we would need for this approach to work is for the family and individual IDs to be the same in the two projects?

[hanars]

To clarify what I stated above, it is impossible to acheive what you want in seqr - either the data will be duplicated in the backend and will have duplicate allele counts, or the data will only be available in one project. There is no way to work around this.

The workflow you describe here is not supported and has undefined behavior. It will either result in a) the data being duplicated to both projects or b) it would load to one project but not the other

[anjalijain22]

Ah I see, got it, thanks for clarifying!