pirl-unc / tsarina / 27371105071
77%

DEFAULT BRANCH: main
Ran
Jobs 1
Files 38
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: 73.88% (+0.09%) from 73.793%
27371105071

push

github

web-flow 
Derive CTA pMHC groups from the cancerdata proteoform registry (15, not 4) (#118)

spanning.py's identical-protein pMHC grouping previously hardcoded 4 groups
(CTAG1A/B, XAGE1A/B, SSX4/SSX4B, MAGEA2/MAGEA2B). It now loads the full
identical-protein CGA registry derived by cancerdata (15 families, incl. the
12-member CT47A family, MAGEA9/9B, SSX2/2B, GAGE12 groups, CT45A2/8/9, ...), so
those families' byte-identical peptides collapse to one proteoform instead of
double-counting across loci.

- tsarina/data/proteoform-groups.csv: materialized mirror of cancerdata's
  registry (deterministic, no runtime dep on cancerdata), regenerated by
  scripts/sync_proteoform_groups.py.
- _load_proteoform_groups() reads the mirror; the four anchor groups are the
  offline fallback if the file is unreadable, and their curated non-symbol
  aliases (NY-ESO-1, GAGED2, XAGE1C-E, SSX4A, MAGEA2A) are overlaid verbatim, so
  anchor resolution is byte-identical. The other 11 families get auto-generated
  member/label aliases.
- test_spanning sync test asserts the mirror matches cancerdata's registry when
  importable; the two dynamic-grouping tests now use synthetic non-registry
  paralogs so they isolate the automatic-peptide-set path.

Single source of truth = cancerdata; this materialization is the interim until
tsarina takes a hard cancerdata dependency. Bumps version to 1.16.0.

2738 of 3706 relevant lines covered (73.88%)

0.74 hits per line

Coverage Regressions

Lines Coverage ∆ File
87
91.03
 -0.02% spanning.py
Jobs
ID Job ID Ran Files Coverage
1 27371105071.1 38
73.88
 GitHub Action Run
Source Files on build 27371105071