generate computed variant id

[reece]

Originally reported by: Reece Hart (Bitbucket: reece, GitHub: reece)

---

Finally close to being able to generate a computed (rather than assigned) variant id for uniqueness.

The idea is to use a hash of a serialized variant as a unique identifier. For example, sha512("<sequence_sha512>:<start>:<end>:<alt_seq>") would generate a hash that uniquely identifies a variant (absent hash collisions).

Such hashes would be useful for:

• Variant equivalence using different sequence aliases (e.g., RefSeq and Ensembl transcripts).
• Querying about the presence of a variant without revealing which variant if it hasn't been already seen.
• Creating an id for variants that are not on an unaccessioned reference sequence (e.g., on a synthetic sequence in a graph sequence representation)

When combined with normalization and other notions of "equivalent" or "representative" variants, this hash would provide a way to declare such relationships without the need for a central authority.

---

• Bitbucket: https://bitbucket.org/biocommons/hgvs/issue/339

[reece]

Original comment by Jerry Liu (Bitbucket: jerryliu2005, GitHub: Unknown):

---

This doesn't seem to handle different genome build. I recently saw a global variant id implementation in this BMC bioinfo. paper (unique for SNVs, deletions, and for insertions/MNVs of up to 2958 inserted nucleotides). Per the article it is unique for SNVs, deletions, and insertions/MNVs of up to 2958 inserted nucleotides. I don't see it is offered as open-source, though. I'm interested in implementing similar ID for our own in-house variant store and would like your input on this. Thx, Jerry

[github-actions[bot]]

This issue is stale because it has been open 90 days with no activity. Remove stale label or comment or this will be closed in 7 days.

[github-actions[bot]]

This issue was closed because it has been stalled for 7 days with no activity.