pirl-unc / hitlist / 24757065298
80%

DEFAULT BRANCH: main
Ran
Jobs 1
Files 21
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: 48.854% (+0.3%) from 48.541%
24757065298

push

github

web-flow 
v1.14.1: add fractionation_ph axis; unpool Tryp-Phos-pH8 / pH10 (#98) (#112)

The two ``Tryp-Phos-pH8`` and ``Tryp-Phos-pH10`` TiO2 arms were
being pooled into a single ``(HeLa, Trypsin/P, 12 fractions, TiO2)``
row bucket with pseudo-replicate labels ``E_pH8`` / ``E_pH10``. That's
wrong: the two runs are complementary high-pH SPE separations at
different buffer pH values, not biological replicates. Peptides elute
differently at pH 8 vs pH 10 because side-chain charge states shift.
``n_replicates_detected = 2`` on the pooled rows was misleading.

## What this adds

- **Row-level ``fractionation_ph`` column** on both peptide and
  protein-level CSVs. ``float``; default ``10.0`` (standard high-pH
  RPLC buffer — 25 mM ammonium formate, pH 10 — used by CCLE, CPTAC,
  HPA, NCI-60, BJ's main fractionation). Only ``Tryp-Phos-pH8`` gets
  ``8.0``; ``Tryp-Phos-pH10`` gets an explicit ``10.0``.

- **Unpooling**: the two Tryp-Phos arms are now distinct rows keyed
  on ``(cell_line, digestion_enzyme, n_fractions_in_run, enrichment,
  fractionation_ph)``. A peptide detected in both pH conditions shows
  up as two rows (one pH 8.0, one pH 10.0) rather than a single row
  with ``n_replicates_detected = 2``. \`abundance_percentile\` is also
  re-ranked PER pH so the two arms' scales are independently comparable.

- **Loader kwargs**: ``load_bulk_peptides(fractionation_ph=...)`` and
  ``load_bulk_proteomics(fractionation_ph=...)`` accept a single float,
  an iterable of floats, or ``None`` (default — no filter).

- **CCLE stamping**: every CCLE row is stamped ``fractionation_ph = 10.0``
  from the source-level default in ``sources.yaml``. BJ rows use the
  row-level value from the ingest script; the source-level default is
  preserved as ``10.0`` for rows without a per-row value.

## Row-count deltas

- Peptide CSV: 2,027,514 → 2,047,003 rows (+19,489 from unpooling)
- Protein CSV: 150,838 → 156,211 rows (+5,373 from unpooling)

The delta comes from peptides/proteins ... (continued)

1535 of 3142 relevant lines covered (48.85%)

0.49 hits per line

Coverage Regressions

Lines Coverage ∆ File
79
50.1
 0.63% builder.py
2
98.56
 -1.44% bulk_proteomics.py
Jobs
ID Job ID Ran Files Coverage
1 24757065298.1 21
48.85
 GitHub Action Run
Source Files on build 24757065298