From 0b12056c4e44bb59448e0535f04c7e7a1a20d009 Mon Sep 17 00:00:00 2001
From: Tim O'Donnell <timodonnell@gmail.com>
Date: Tue, 7 Jan 2020 18:26:58 -0500
Subject: [PATCH] fix
---
.../models_class1_pan_variants/GENERATE.sh | 12 ++++--------
.../reassign_mass_spec_training_data.py | 13 ++++++++-----
2 files changed, 12 insertions(+), 13 deletions(-)
diff --git a/downloads-generation/models_class1_pan_variants/GENERATE.sh b/downloads-generation/models_class1_pan_variants/GENERATE.sh
index 13d7aecb..7782d2fa 100755
--- a/downloads-generation/models_class1_pan_variants/GENERATE.sh
+++ b/downloads-generation/models_class1_pan_variants/GENERATE.sh
@@ -69,8 +69,8 @@ then
python generate_hyperparameters.py hyperparameters.production.yaml compact_peptide > hyperparameters.compact_peptide.yaml
fi
-#VARIANTS=( no_additional_ms_ms_only_0nm ms_only_0nm no_additional_ms_0nm 0nm no_additional_ms ms_only no_pretrain compact_peptide 34mer_sequence single_hidden_no_pretrain affinity_only )
-VARIANTS=( no_additional_ms_ms_only_0nm ms_only_0nm no_additional_ms_0nm 0nm no_additional_ms ms_only no_pretrain compact_peptide 34mer_sequence )
+#VARIANTS=( no_additional_ms_ms_only_0nm ms_only_0nm no_additional_ms_0nm 0nm no_additional_ms no_pretrain compact_peptide 34mer_sequence single_hidden_no_pretrain affinity_only )
+VARIANTS=( no_additional_ms_ms_only_0nm ms_only_0nm no_additional_ms_0nm 0nm no_additional_ms no_pretrain compact_peptide 34mer_sequence )
for kind in "${VARIANTS[@]}"
do
@@ -102,6 +102,7 @@ do
python reassign_mass_spec_training_data.py \
"$(mhcflurry-downloads path data_curated)/curated_training_data.no_additional_ms.csv.bz2" \
--set-measurement-value 0 \
+ --drop-negative-ms \
--ms-only \
--out-csv "$TRAINING_DATA"
HYPERPARAMETERS=hyperparameters.production.yaml
@@ -133,16 +134,11 @@ do
python reassign_mass_spec_training_data.py \
"$(mhcflurry-downloads path data_curated)/curated_training_data.mass_spec.csv.bz2" \
--set-measurement-value 0 \
+ --drop-negative-ms \
--out-csv "$TRAINING_DATA"
HYPERPARAMETERS=hyperparameters.production.yaml
fi
- if [ "$kind" == "ms_only" ]
- then
- TRAINING_DATA="$(mhcflurry-downloads path data_curated)/curated_training_data.mass_spec.csv.bz2"
- HYPERPARAMETERS=hyperparameters.production.yaml
- fi
-
if [ "$kind" == "affinity_only" ]
then
TRAINING_DATA="$(mhcflurry-downloads path data_curated)/curated_training_data.affinity.csv.bz2"
diff --git a/downloads-generation/models_class1_pan_variants/reassign_mass_spec_training_data.py b/downloads-generation/models_class1_pan_variants/reassign_mass_spec_training_data.py
index 15838419..185eb75f 100644
--- a/downloads-generation/models_class1_pan_variants/reassign_mass_spec_training_data.py
+++ b/downloads-generation/models_class1_pan_variants/reassign_mass_spec_training_data.py
@@ -11,6 +11,7 @@ parser = argparse.ArgumentParser(usage=__doc__)
parser.add_argument("data", metavar="CSV", help="Training data")
parser.add_argument("--ms-only", action="store_true", default=False)
+parser.add_argument("--drop-negative-ms", action="store_true", default=False)
parser.add_argument("--set-measurement-value", type=float)
parser.add_argument("--out-csv")
@@ -21,11 +22,13 @@ def go(args):
df = pandas.read_csv(args.data)
print(df)
- bad = df.loc[
- (df.measurement_kind == "mass_spec") &
- (df.measurement_inequality != "<")
- ]
- assert len(bad) == 0, bad
+ if args.drop_negative_ms:
+ bad = df.loc[
+ (df.measurement_kind == "mass_spec") &
+ (df.measurement_inequality != "<")
+ ]
+ print("Dropping ", len(bad))
+ df = df.loc[~df.index.isin(bad.index)].copy()
if args.ms_only:
print("Filtering to MS only")
--
GitLab