From b50dc80c6ec0874a0b4a61873506847ae9e3471b Mon Sep 17 00:00:00 2001
From: Tim O'Donnell <timodonnell@gmail.com>
Date: Sun, 15 Sep 2019 21:40:41 -0400
Subject: [PATCH] Drop insufficiently specific allele names like 'HLA-A03' in
data curation script. Fixes #119
---
downloads-generation/data_curated/curate.py | 12 ++++++++++++
1 file changed, 12 insertions(+)
diff --git a/downloads-generation/data_curated/curate.py b/downloads-generation/data_curated/curate.py
index 08994ee3..5f99ccd5 100755
--- a/downloads-generation/data_curated/curate.py
+++ b/downloads-generation/data_curated/curate.py
@@ -134,6 +134,18 @@ def load_data_iedb(iedb_csv, include_qualitative=True, include_mass_spec=False):
(~iedb_df["Allele Name"].str.contains("CD1"))
]
+ # Drop insufficiently specific allele names like "HLA-A03":
+ insuffient_mask = (
+ (~iedb_df["Allele Name"].str.upper().str.startswith("H2-")) &
+ (~iedb_df["Allele Name"].str.upper().str.startswith("H-2-")) &
+ (~iedb_df["Allele Name"].str.upper().str.startswith("MAMU")) &
+ (iedb_df["Allele Name"].str.findall("[0-9]").str.len() < 4)
+ )
+ print("Dropping %d records with insufficiently-specific allele names:" %
+ insuffient_mask.sum())
+ print(iedb_df.loc[insuffient_mask]["Allele Name"].value_counts())
+ iedb_df = iedb_df.loc[~insuffient_mask]
+
iedb_df["allele"] = iedb_df["Allele Name"].map(normalize_allele_name)
print("Dropping un-parseable alleles: %s" % ", ".join(
iedb_df.loc[iedb_df.allele == "UNKNOWN"]["Allele Name"].unique()))
--
GitLab