Open frederikziebell opened 3 years ago
Hi,
Yes, donor deconvolution will not be possible any more. I would say the strategy to use random bases is suboptimal. The fact that you do find mismatches to the reference in certain locations is already carrying information content that could be used to infer information on the donor. My suggestion would be that authors of such data should provide the anonymized data along with the correct metadata to allow for accounting of donor effects.
Best, Christoph
I haven't tried the tool but from reading the paper it shouldn't be possible anymore to demultiplex donors, e.g. by using Vireo. This would be a major drawback because one could not account for donor effects in the downstream analysis.
If instead of always substituting the reference base, one would randomly select a base for each (donor, SNP) pair, donors are anonymous but could still be demultiplexed. Does this make sense?