# How to anonymise a bam file

I have a bam file of, for instance, RNA-Seq, which contains patient-identifiable data in the form of Single Nucleotide Polymorphisms (SNPs) throughout.

I would like a technique to take this aligned bam file and replace all the non-reference bases with the reference bases at the equivalent positions (hard-clipping any soft-clipped bases while doing so), in order to have an anonymised bam file without the patient-identifiable information.

What is the best way to do this? I want the resulting anonymised bam file to have the reads in the same positions, for downstream tools that care about the alignment positions.

• Hi @KraZug normally patient confidentiality is maintained through anonminity of name and personal information, but not to anonymize the DNA by ad hoc editting. No ethics committee would request that. Personally, I don't agree here.
– M__
Jun 19 '20 at 20:51
• In the event of a publication, normally you can submit to ega or dbgap ncbi.nlm.nih.gov/gap/docs/submissionguide/…. and they have a set of guidelines on how u de-identify samples Jun 19 '20 at 21:49

You can simply replace the sequence and quality scores with a single * in each field. The CIGAR information will still be there, so it will be apparent where there were InDels and possibly mismatches (assuming your aligning uses = and X CIGAR operations), but since the actual sequence of these variants cannot be known this is quite anonymized.