pirl-unc / tsarina / 25516187047
70%

DEFAULT BRANCH: main
Ran
Jobs 1
Files 35
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 69.728% (+0.07%) from 69.658%
25516187047

push

github

web-flow 
v1.2.21: handle peptide_attribution provenance from hitlist v1.30.39 (#62) (#64)

* v1.2.21: handle peptide_attribution provenance from hitlist v1.30.39 (#62)

Hitlist v1.30.39 (pirl-unc/hitlist#231 closing #45) added a new
``mhc_allele_provenance`` value: ``peptide_attribution``.  It's
emitted for class-only IEDB rows whose peptide appears in a paper-
supplement-derived attribution CSV — the candidate-allele set has
been narrowed from the disease-wide donor union to the matched donor's
typed alleles.  Sarkizova 2020's 36 K patient-tumor MS rows
(PMID 31844290) all flip from class-only ``"HLA class I"`` to
``"peptide_attribution"`` provenance after the next corpus rebuild.

``peptide_attribution`` is *strictly more specific* than
``sample_allele_match``: same shape (multi-allele set in
``mhc_allele_set``) but narrowed via the paper's per-peptide → per-
sample mapping rather than a row-level ``Host | MHC Types Present``.

Two ``spanning.py`` filters previously checked
``provenance == "sample_allele_match"`` literally and silently
skipped ``peptide_attribution`` rows:

- ``_extend_alleles_from_observations`` (line 1665): wouldn't extend
  the panel allele list with peptide-attributed donor sets, so
  ``best-of-sample`` scoring couldn't see the row's alleles.
- ``_attribute_panel_evidence_with_observations`` (line 1841): wouldn't
  attribute panel evidence to peptide-attributed rows under
  ``sample_allele_ms``.

Fix: factor a module-level ``_SAMPLE_NARROWED_PROVENANCES`` frozenset
that holds both values, and check membership in both call sites.
Treats ``peptide_attribution`` and ``sample_allele_match`` identically
since they share data shape and downstream semantics — only the
narrowing source differs.

Tests:

- ``test_peptide_attribution_provenance_treated_like_sample_allele_match``:
  Sarkizova-shaped row gets ``sample_allele_ms`` evidence tier.
- ``test_extend_alleles_picks_up_peptide_attribution_donor_set``:
  best-of-sample picks the right haploty... (continued)

2285 of 3277 relevant lines covered (69.73%)

0.7 hits per line

Coverage Regressions

Lines Coverage ∆ File
85
90.52
 0.22% spanning.py
Jobs
ID Job ID Ran Files Coverage
1 25516187047.1 35
69.73
 GitHub Action Run
Source Files on build 25516187047