From 6102b31e86c0e445e28a76fdcd98f1e7f4e85c22 Mon Sep 17 00:00:00 2001
From: Tim O'Donnell <timodonnell@gmail.com>
Date: Mon, 24 Sep 2018 15:20:12 -0400
Subject: [PATCH] fix
---
mhcflurry/allele_encoding.py | 22 +++++++++++++---------
mhcflurry/allele_encoding_transforms.py | 6 ++++--
mhcflurry/class1_affinity_predictor.py | 18 +++++++++++++-----
test/test_allele_encoding.py | 3 ++-
4 files changed, 32 insertions(+), 17 deletions(-)
diff --git a/mhcflurry/allele_encoding.py b/mhcflurry/allele_encoding.py
index e7d47cf0..35a124db 100644
--- a/mhcflurry/allele_encoding.py
+++ b/mhcflurry/allele_encoding.py
@@ -74,13 +74,16 @@ class AlleleEncoding(object):
"allele_representations",
encoding_name)
if cache_key not in self.encoding_cache:
- if callable(encoding_name):
- vector_encoded = encoding_name(self)
- assert len(vector_encoded)== len(self.allele_to_sequence)
- elif ":" in encoding_name:
- # Apply transform
- (transform_name, rest) = encoding_name.split(":", 2)
- preliminary_encoded = self.allele_representations(rest)
+ if ":" in encoding_name:
+ # Transform
+ pieces = encoding_name.split(":", 3)
+ if pieces[0] != "transform":
+ raise RuntimeError(
+ "Expected 'transform' but saw: %s" % pieces[0])
+ if len(pieces) == 1:
+ raise RuntimeError("Expected: 'transform:<name>[:argument]")
+ transform_name = pieces[1]
+ argument = None if len(pieces) == 2 else pieces[2]
try:
transform = self.transforms[transform_name]
except KeyError:
@@ -88,8 +91,9 @@ class AlleleEncoding(object):
"Unsupported transform: %s. Supported transforms: %s" % (
transform_name,
" ".join(self.transforms) if self.transforms else "(none)"))
-
- vector_encoded = transform.transform(preliminary_encoded)
+ vector_encoded = (
+ transform.transform(self) if argument is None
+ else transform.transform(self, argument))
else:
# No transform.
index_encoded_matrix = amino_acid.index_encoding(
diff --git a/mhcflurry/allele_encoding_transforms.py b/mhcflurry/allele_encoding_transforms.py
index 88b361f5..018d831e 100644
--- a/mhcflurry/allele_encoding_transforms.py
+++ b/mhcflurry/allele_encoding_transforms.py
@@ -5,7 +5,7 @@ import sklearn.decomposition
class AlleleEncodingTransform(object):
- def transform(self, data):
+ def transform(self, allele_encoding, argument=None):
raise NotImplementedError()
def get_fit(self):
@@ -61,7 +61,9 @@ class PCATransform(AlleleEncodingTransform):
self.model.mean_ = fit["mean"]
self.model.components_ = fit["components"]
- def transform(self, allele_representations):
+ def transform(self, allele_encoding, underlying_representation):
+ allele_representations = allele_encoding.allele_representations(
+ underlying_representation)
if not self.is_fit():
self.fit(allele_representations)
flattened = allele_representations.reshape(
diff --git a/mhcflurry/class1_affinity_predictor.py b/mhcflurry/class1_affinity_predictor.py
index 1e9ad71f..87c00eda 100644
--- a/mhcflurry/class1_affinity_predictor.py
+++ b/mhcflurry/class1_affinity_predictor.py
@@ -816,7 +816,8 @@ class Class1AffinityPredictor(object):
alleles=None,
allele=None,
throw=True,
- centrality_measure=DEFAULT_CENTRALITY_MEASURE):
+ centrality_measure=DEFAULT_CENTRALITY_MEASURE,
+ model_kwargs={}):
"""
Predict nM binding affinities.
@@ -840,6 +841,8 @@ class Class1AffinityPredictor(object):
centrality_measure : string or callable
Measure of central tendency to use to combine predictions in the
ensemble. Options include: mean, median, robust_mean.
+ model_kwargs : dict
+ Additional keyword arguments to pass to Class1NeuralNetwork.predict
Returns
-------
@@ -853,6 +856,7 @@ class Class1AffinityPredictor(object):
include_percentile_ranks=False,
include_confidence_intervals=False,
centrality_measure=centrality_measure,
+ model_kwargs=model_kwargs
)
return df.prediction.values
@@ -865,7 +869,8 @@ class Class1AffinityPredictor(object):
include_individual_model_predictions=False,
include_percentile_ranks=True,
include_confidence_intervals=True,
- centrality_measure=DEFAULT_CENTRALITY_MEASURE):
+ centrality_measure=DEFAULT_CENTRALITY_MEASURE,
+ model_kwargs={}):
"""
Predict nM binding affinities. Gives more detailed output than `predict`
method, including 5-95% prediction intervals.
@@ -897,6 +902,8 @@ class Class1AffinityPredictor(object):
centrality_measure : string or callable
Measure of central tendency to use to combine predictions in the
ensemble. Options include: mean, median, robust_mean.
+ model_kwargs : dict
+ Additional keyword arguments to pass to Class1NeuralNetwork.predict
Returns
-------
@@ -1002,7 +1009,8 @@ class Class1AffinityPredictor(object):
for (i, model) in enumerate(self.class1_pan_allele_models):
predictions_array[mask, i] = model.predict(
masked_peptides,
- allele_encoding=masked_allele_encoding)
+ allele_encoding=masked_allele_encoding,
+ **model_kwargs)
if self.allele_to_allele_specific_models:
unsupported_alleles = [
@@ -1031,7 +1039,7 @@ class Class1AffinityPredictor(object):
# Common case optimization
for (i, model) in enumerate(models):
predictions_array[:, num_pan_models + i] = (
- model.predict(peptides))
+ model.predict(peptides, **model_kwargs))
elif mask.sum() > 0:
peptides_for_allele = EncodableSequences.create(
df.ix[mask].peptide.values)
@@ -1039,7 +1047,7 @@ class Class1AffinityPredictor(object):
predictions_array[
mask,
num_pan_models + i,
- ] = model.predict(peptides_for_allele)
+ ] = model.predict(peptides_for_allele, **model_kwargs)
if callable(centrality_measure):
centrality_function = centrality_measure
diff --git a/test/test_allele_encoding.py b/test/test_allele_encoding.py
index ef526e2d..a6c1a972 100644
--- a/test/test_allele_encoding.py
+++ b/test/test_allele_encoding.py
@@ -47,7 +47,8 @@ def test_pca():
"A*02:03": "AE",
}
)
- encoded1 = encoding.fixed_length_vector_encoded_sequences("pca:BLOSUM62")
+ encoded1 = encoding.fixed_length_vector_encoded_sequences(
+ "transform:pca:BLOSUM62")
numpy.testing.assert_array_equal(encoded1[0], encoded1[2])
assert not numpy.array_equal(encoded1[0], encoded1[1])
--
GitLab