When a user provides an MD5 that doesn't match any catalog entry (e.g. 3a7b8f4c8f5a975a9e656ea77bfa1498 for chr1), it may be from a modified reference (repeat-masked, soft-masked via sdust, etc.) rather than an error.
We could pre-compute and catalog MD5s for common FASTA mutations:
- RepeatMasker (hard-masked)
- sdust / DUST (soft-masked)
- Other common transformations (e.g. lowercase conversion)
This would allow ref-solver to identify modified references and report what transformation was likely applied.
When a user provides an MD5 that doesn't match any catalog entry (e.g.
3a7b8f4c8f5a975a9e656ea77bfa1498for chr1), it may be from a modified reference (repeat-masked, soft-masked via sdust, etc.) rather than an error.We could pre-compute and catalog MD5s for common FASTA mutations:
This would allow ref-solver to identify modified references and report what transformation was likely applied.