pirl-unc / hitlist / 24101283174
44%

DEFAULT BRANCH: main
Ran
Jobs 1
Files 13
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: 32.883% (-4.9%) from 37.789%
24101283174

push

github

web-flow 
Audit PMID overrides: fix metadata, add per-sample conditions (#6)

* Audit and fix PMID override metadata against PubMed papers

- Fix swapped entries: 28228285 is Abelin 2017 (16 alleles),
  31844290 is Sarkizova 2020 (95 alleles)
- Fix wrong attributions: 26992070 is Ritz 2016 (not Caron 2015),
  35051231 is Nicholas 2022 (not Pyke 2022), 27862975 is Ritz 2017
  (not Bassani-Sternberg), 31092671 is Guasp 2019 (not Alvarez 2020)
- Correct all paper titles to match actual PubMed entries
- Add conditional rules for mixed studies (26992070, 31154438,
  29786170) to properly classify healthy controls vs cancer samples
- Add perturbations metadata: ERAP1/ERAP2 CRISPR KOs, influenza
  infection, DC/T/B cell activation protocols, moDC differentiation
- Add hla_alleles metadata for 12 of 14 studies
- Update docs/pmid-curation.md to reflect corrections

* Refine PMID sample-level detail and fix perturbation metadata

- 36589698: Remove spurious healthy override rule — no healthy donor
  MS data exists in IEDB (29 elution records, all cancer). Add
  structured sample inventory (4 cell lines + 8 ovarian tumors).
- 38920720: Fix perturbations — MS was on peptide-pulsed moDCs with
  IFN-gamma + LPS maturation (not mRNA electroporation, which was
  T cell assays only).
- 35051231: Clarify ALL MS lung samples were influenza-infected,
  no uninfected controls profiled by MS.
- 31844290: Add IFN-gamma perturbation note for 6 validation tumor
  cell lines (not in mono-allelic 721.221 training data).

* Audit PMID overrides: fix metadata, add per-sample conditions, bump to 0.8.1

Verified all 14 PMID entries against actual PubMed papers:

Corrections:
- Fix swapped entries: 28228285=Abelin 2017, 31844290=Sarkizova 2020
- Fix wrong attributions: 26992070=Ritz 2016 (not Caron 2015),
  35051231=Nicholas 2022 (not Pyke 2022), 27862975=Ritz 2017
  (not Bassani-Sternberg), 31092671=Guasp 2019 (not Alvarez 2020)
- Correct all paper titles to match PubMed
- 36589698: remove s... (continued)

438 of 1332 relevant lines covered (32.88%)

0.33 hits per line

Coverage Regressions

Lines Coverage ∆ File
98
0.0
 0.0% cli.py
2
73.63
 -1.0% curation.py
Jobs
ID Job ID Ran Files Coverage
1 24101283174.1 13
32.88
 GitHub Action Run
Source Files on build 24101283174