Fix 3'-rule errors for certain sequence patterns #75
Conversation
Codecov Report
@@ Coverage Diff @@
## master #75 +/- ##
==========================================
- Coverage 45.63% 45.60% -0.03%
==========================================
Files 16 16
Lines 2003 2004 +1
Branches 64 64
==========================================
Hits 914 914
- Misses 1025 1026 +1
Partials 64 64
|
d8b3818 to
b4be611
Compare
b4be611 to
c614b8e
Compare
| (if (= alt \*) | ||
| (protein-substitution (+ ppos offset) (str ref) (str alt)) ; eventually fs-ter-substitution |
There was a problem hiding this comment.
📝 In certain cases, frameshift would be fixed as fs-ter-substitution by the 3'-rule application to protein sequence. So I have added the fs-ter-substition check here.
There was a problem hiding this comment.
Indeed, such cases potentially occur as you said. It is extremely rare.
Do you have any case of that? I would like you to add the test case if you already have.
No problem even if you do not have the case. I will approve this PR even then because I think it is a bit difficult to prepare the test case.
There was a problem hiding this comment.
Thanks for pointing this out! I think my REPL history still contains the details. I will try to retrieve it.
There was a problem hiding this comment.
Sorry, the variant query I attempted appears no longer comes in this path by the backward-shift fix. 🤔
Instead, I have added test cases for varity.vcf-to-hgvs.protein/mutation in 1854562, which includes inputs before and after the 3'-rule of coding DNA. The 3'-rule for protein sequence and fs-ter-substitution check work correctly in both cases.
There was a problem hiding this comment.
Okay, thank you very much. The test case using varity.vcf-to-hgvs.protein/mutation looks good to me.
| (if (= alt \*) | ||
| (protein-substitution (+ ppos offset) (str ref) (str alt)) ; eventually fs-ter-substitution |
There was a problem hiding this comment.
Indeed, such cases potentially occur as you said. It is extremely rare.
Do you have any case of that? I would like you to add the test case if you already have.
No problem even if you do not have the case. I will approve this PR even then because I think it is a bit difficult to prepare the test case.
6b2ef4c to
1854562
Compare
| (if (= alt \*) | ||
| (protein-substitution (+ ppos offset) (str ref) (str alt)) ; eventually fs-ter-substitution |
There was a problem hiding this comment.
Okay, thank you very much. The test case using varity.vcf-to-hgvs.protein/mutation looks good to me.
Resolve #72 and #74.
This PR addresses the problem of 3' shifting reported in issue #74 and the out-of-bounds exception reported in issue #72.
I have made some refinements to the
varity.vcf-to-hgvs. Specifically, I have fixedvarity.vcf-to-hgvs.common/backward-shiftfunction to work correctly for certain sequence patterns and added error handling tovarity.vcf-to-hgvs.common/apply-3'-rulefor rare cases.Additionally, I have fixed the exon boundary handling in
varity.vcf-to-hgvs.protein/read-sequence-info, which was implemented in #68.Added test cases is originates from #71 and #73 by @nokara26 Thanks! ❤️