From d1ef4aae5a49093a0fb82935f98f75a7140ae217 Mon Sep 17 00:00:00 2001
From: Tim O'Donnell <timodonnell@gmail.com>
Date: Thu, 16 May 2019 09:51:19 -0400
Subject: [PATCH] add test
---
downloads-generation/data_curated/curate.py | 2 +-
mhcflurry/downloads.yml | 2 +-
mhcflurry/train_pan_allele_models_command.py | 6 --
test/test_class1_pan.py | 89 ++++++++++++++++++++
4 files changed, 91 insertions(+), 8 deletions(-)
create mode 100644 test/test_class1_pan.py
diff --git a/downloads-generation/data_curated/curate.py b/downloads-generation/data_curated/curate.py
index d61719f7..2c5f0319 100755
--- a/downloads-generation/data_curated/curate.py
+++ b/downloads-generation/data_curated/curate.py
@@ -145,7 +145,7 @@ def load_data_iedb(iedb_csv, include_qualitative=True, include_mass_spec=False):
quantitative["measurement_type"] = "quantitative"
quantitative["measurement_inequality"] = quantitative[
"Measurement Inequality"
- ].fillna("=")
+ ].fillna("=").map(lambda s: {">=": ">", "<=": "<"}.get(s, s))
print("Quantitative measurements: %d" % len(quantitative))
qualitative = iedb_df.ix[iedb_df["Units"].isnull()].copy()
diff --git a/mhcflurry/downloads.yml b/mhcflurry/downloads.yml
index f1f3e980..7b87201e 100644
--- a/mhcflurry/downloads.yml
+++ b/mhcflurry/downloads.yml
@@ -41,7 +41,7 @@ releases:
default: false
- name: data_curated
- url: https://github.com/openvax/mhcflurry/releases/download/pan-dev1/data_curated.20190514.tar.bz2
+ url: https://github.com/openvax/mhcflurry/releases/download/pan-dev1/data_curated.20190516.tar.bz2
default: true
# Older downloads
diff --git a/mhcflurry/train_pan_allele_models_command.py b/mhcflurry/train_pan_allele_models_command.py
index bc9bb9ab..1520502a 100644
--- a/mhcflurry/train_pan_allele_models_command.py
+++ b/mhcflurry/train_pan_allele_models_command.py
@@ -246,12 +246,6 @@ def main(args):
print("Data inequalities:")
print(df.measurement_inequality.value_counts())
- df.measurement_inequality = df.measurement_inequality.map(lambda s: {
- ">=": ">", "<=": "<"
- }.get(s, s))
-
- print("Data inequalities after adjustment:")
- print(df.measurement_inequality.value_counts())
if args.ignore_inequalities and "measurement_inequality" in df.columns:
print("Dropping measurement_inequality column")
diff --git a/test/test_class1_pan.py b/test/test_class1_pan.py
new file mode 100644
index 00000000..acea90a2
--- /dev/null
+++ b/test/test_class1_pan.py
@@ -0,0 +1,89 @@
+"""
+Tests for training and predicting using Class1 pan-allele models.
+"""
+
+import json
+import os
+import shutil
+import tempfile
+import subprocess
+from copy import deepcopy
+
+import pandas
+
+from numpy.testing import assert_array_less, assert_equal
+
+from mhcflurry import Class1AffinityPredictor,Class1NeuralNetwork
+from mhcflurry.allele_encoding import AlleleEncoding
+from mhcflurry.downloads import get_path
+
+
+HYPERPARAMETERS = {
+ 'activation': 'tanh',
+ 'allele_dense_layer_sizes': [],
+ 'batch_normalization': False,
+ 'dense_layer_l1_regularization': 0.0,
+ 'dense_layer_l2_regularization': 0.0,
+ 'dropout_probability': 0.5,
+ 'early_stopping': True,
+ 'init': 'glorot_uniform',
+ 'layer_sizes': [64],
+ 'learning_rate': None,
+ 'locally_connected_layers': [],
+ 'loss': 'custom:mse_with_inequalities',
+ 'max_epochs': 5000,
+ 'minibatch_size': 128,
+ 'optimizer': 'rmsprop',
+ 'output_activation': 'sigmoid',
+ 'patience': 20,
+ 'peptide_allele_merge_activation': '',
+ 'peptide_allele_merge_method': 'concatenate',
+ 'peptide_amino_acid_encoding': 'BLOSUM62',
+ 'peptide_dense_layer_sizes': [],
+ 'peptide_encoding': {
+ 'alignment_method': 'left_pad_centered_right_pad',
+ 'max_length': 15,
+ 'vector_encoding_name': 'BLOSUM62',
+ },
+ 'random_negative_affinity_max': 50000.0,
+ 'random_negative_affinity_min': 20000.0,
+ 'random_negative_constant': 25,
+ 'random_negative_distribution_smoothing': 0.0,
+ 'random_negative_match_distribution': True,
+ 'random_negative_rate': 0.2,
+ 'train_data': {},
+ 'validation_split': 0.1,
+}
+
+
+ALLELE_TO_SEQUENCE = pandas.read_csv(
+ get_path(
+ "allele_sequences", "allele_sequences.csv"),
+ index_col=0).sequence.to_dict()
+
+
+TRAIN_DF = pandas.read_csv(
+ get_path(
+ "data_curated", "curated_training_data.no_mass_spec.csv.bz2"))
+
+TRAIN_DF = TRAIN_DF.loc[TRAIN_DF.allele.isin(ALLELE_TO_SEQUENCE)]
+TRAIN_DF = TRAIN_DF.loc[TRAIN_DF.peptide.str.len() >= 8]
+TRAIN_DF = TRAIN_DF.loc[TRAIN_DF.peptide.str.len() <= 15]
+
+
+def test_train_simple():
+ network = Class1NeuralNetwork(**HYPERPARAMETERS)
+ allele_encoding = AlleleEncoding(
+ TRAIN_DF.allele.values,
+ allele_to_sequence=ALLELE_TO_SEQUENCE)
+ network.fit(
+ TRAIN_DF.peptide.values,
+ affinities=TRAIN_DF.measurement_value.values,
+ allele_encoding=allele_encoding,
+ inequalities=TRAIN_DF.measurement_inequality.values)
+
+ predictions = network.predict(
+ peptides=TRAIN_DF.peptide.values,
+ allele_encoding=allele_encoding)
+
+ print(pandas.Series(predictions).describe())
--
GitLab