From ba9360750a7481cf8345c663cdb7c22b944da7f6 Mon Sep 17 00:00:00 2001
From: Tim O'Donnell <timodonnell@gmail.com>
Date: Sat, 18 Jan 2020 17:09:03 -0500
Subject: [PATCH] fix
---
downloads-generation/models_class1_cleavage/GENERATE.sh | 2 +-
.../models_class1_cleavage/make_train_data.py | 7 ++++---
2 files changed, 5 insertions(+), 4 deletions(-)
diff --git a/downloads-generation/models_class1_cleavage/GENERATE.sh b/downloads-generation/models_class1_cleavage/GENERATE.sh
index 730ed339..2a27b160 100755
--- a/downloads-generation/models_class1_cleavage/GENERATE.sh
+++ b/downloads-generation/models_class1_cleavage/GENERATE.sh
@@ -95,7 +95,7 @@ else
--predictions "$(mhcflurry-downloads path data_mass_spec_benchmark)/predictions/all.mhcflurry.combined" \
--proteome-peptides "$(mhcflurry-downloads path data_mass_spec_benchmark)/proteome_peptides.all.csv.bz2" \
--ppv-multiplier 100 \
- --hit-multiplier-to-take 1 \
+ --hit-multiplier-to-take 2 \
--out "$(pwd)/train_data.csv"
bzip2 -f train_data.csv
fi
diff --git a/downloads-generation/models_class1_cleavage/make_train_data.py b/downloads-generation/models_class1_cleavage/make_train_data.py
index 0b3e361c..9200f824 100644
--- a/downloads-generation/models_class1_cleavage/make_train_data.py
+++ b/downloads-generation/models_class1_cleavage/make_train_data.py
@@ -80,9 +80,7 @@ def load_predictions(dirname, result_df=None, columns=None):
manifest_df = manifest_df.loc[manifest_df.col.isin(result_df.columns)]
- print("Will load", len(peptides), "peptides and", len(manifest_df), "cols")
-
- for _, row in tqdm.tqdm(manifest_df.iterrows(), total=len(manifest_df)):
+ for _, row in manifest_df.iterrows():
with open(os.path.join(dirname, row.path), "rb") as fd:
value = numpy.load(fd)['arr_0']
if mask is not None:
@@ -109,6 +107,9 @@ def run():
print("Subselected to %d monoallelic samples" % hit_df.sample_id.nunique())
hit_df["allele"] = hit_df.hla
+ hit_df = hit_df.loc[hit_df.allele.str.match("^HLA-[ABC]")]
+ print("Subselected to %d HLA-A/B/C samples" % hit_df.sample_id.nunique())
+
if args.exclude_contig:
new_hit_df = hit_df.loc[
hit_df.protein_primary_ensembl_contig.astype(str) !=
--
GitLab