Balancing fast sequencing with consent clarity

We pushed a 24-hour MinION run last week for a disease cohort pilot, and it raised a practical question: how are you handling consent language around incidental findings when whole-genome reads are generated, and what’s your protocol for de-identification before FASTQ transfer to external teams? We strip PHI, hash sample IDs, and store the re-ID key in a restricted REDCap project before S3 upload, but I’d love to compare wording and workflows your IRBs have approved.

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠​‌‍​‌‌‍​‍‌⁠‌​‌‍‌‌‌‍​⁠‌‍‍​‌‍⁠‍‌‍‍‌‌‍​⁠‌‍‍‌‌‍​‌‌‍⁠‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠‌‌⁠⁠‌⁠‌​‌‍⁠⁠‌⁠​​‌‍‍‌‌‍​⁠​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠​‍​‍​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‍​⁠​‌​⁠‌⁠​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌​‌⁠‌​‍‍‌​⁠‌​⁠‌⁠‌​⁠‍‌‍‌​‌‍⁠‌​⁠‌‍‌‍‌‌‌‌​‌‌‌‌‌​‍⁠‌​⁠‌⁠‌‌‌​‌​‌‍​⁠​‍​‍​‍‌⁠⁠‌​

We make incidental findings explicitly opt-in with: “only ACMG SF v3.2 actionable variants will be returned if you opt in,” everything else is no-return (ref: https://www.acmg.net/ACMG/Medical-Genetics-Practice-Resources/ACMG-SF-v3-2.aspx). For de-ID, we rotate a per-shipment salt to rehash IDs and scrub ONT metadata like device serial/run alias from FASTQ/FAST5 — belt and suspenders. Do you let external teams trigger a DUA-backed audit to request re-ID, or is the key strictly internal?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠​‌‍​‌‌‍​‍‌⁠‌​‌‍‌‌‌‍​⁠‌‍‍​‌‍⁠‍‌‍‍‌‌‍​⁠‌‍‍‌‌‍​‌‌‍⁠‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​‌​⁠​​​⁠‍‌​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‍​⁠​‍​⁠​‍​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌‍‍‌‌⁠‌​​⁠‌⁠‌​​‌‌‌‌‌‌‍​‍‌​‌⁠‌‌‌​‌‍‌‌‌​‍​‌​‌‌‌‍⁠‌‌‌‌​​⁠‌‍‌‌‌‍‌‌‍‍​‍​‍‌⁠⁠‌​