Merge pull request #24 from jisraeli/master

architectures test with random sequences and labels; updated dependencies

README.md

@@ -11,7 +11,7 @@ To install the latest released version of DragoNN, install the [Anaconda](https:
 ```
 conda install dragonn -c kundajelab
 ```
-DragoNN is compatible with Python2 and Python3. Specific optional features such as [DeepLIFT](https://github.com/kundajelab/deeplift) method and [MOE](https://github.com/Yelp/MOE) are compatible with Python2 only.
+DragoNN is compatible with Python2 and Python3. Specific optional features such as [DeepLIFT](https://github.com/kundajelab/deeplift) and [MOE](https://github.com/Yelp/MOE) are compatible with Python2 only.
 
 
 ## 15 seconds to your first DragoNN model

setup.py

@@ -7,10 +7,10 @@
     'version': '0.1.2',
     'packages': ['dragonn'],
     'setup_requires': [],
-    'install_requires': ['numpy>=1.9', 'keras==0.3.3', 'deeplift==0.3', 'shapely', 'simdna==0.1', 'matplotlib<=1.5.3',
+    'install_requires': ['numpy>=1.9', 'keras==0.3.3', 'deeplift==0.3', 'shapely', 'simdna==0.2', 'matplotlib<=1.5.3',
                          'scikit-learn', 'pydot_ng==1.0.0', 'h5py'],
     'dependency_links': ["https://github.com/kundajelab/deeplift/tarball/v0.3-alpha#egg=deeplift-0.3",
-                         "https://github.com/kundajelab/simdna/tarball/0.1#egg=simdna-0.1"],
+                         "https://github.com/kundajelab/simdna/tarball/0.2#egg=simdna-0.2"],
     'scripts': [],
     'entry_points': {'console_scripts': ['dragonn = dragonn.__main__:main']},
     'name': 'dragonn'

tests/test_architectures.py

@@ -6,25 +6,19 @@
 from collections import OrderedDict
 from dragonn.models import SequenceDNN
 from dragonn.utils import one_hot_encode, reverse_complement
-from simdna.simulations import simulate_single_motif_detection
 try:
     from sklearn.model_selection import train_test_split  # sklearn >= 0.18
 except ImportError:
     from sklearn.cross_validation import train_test_split  # sklearn < 0.18
 
 
-def run(use_deep_CNN, use_RNN, label, golden_first_sequence, golden_results):
+def run(use_deep_CNN, use_RNN, label, golden_results):
     seq_length = 100
     num_sequences = 200
-    num_positives = 100
-    num_negatives = num_sequences - num_positives
-    GC_fraction = 0.4
     test_fraction = 0.2
     num_epochs = 1
-    sequences, labels, embeddings = simulate_single_motif_detection(
-        'SPI1_disc1', seq_length, num_positives, num_negatives, GC_fraction)
-    assert sequences[0] == golden_first_sequence, 'first sequence = {}, golden = {}'.format(
-        sequences[0], golden_first_sequence)
+    sequences = np.array([''.join(random.choice('ACGT') for base in range(seq_length)) for sequence in range(num_sequences)])
+    labels = np.random.choice((True, False), size=num_sequences)[:, None]
     encoded_sequences = one_hot_encode(sequences)
     X_train, X_test, y_train, y_test = train_test_split(
         encoded_sequences, labels, test_size=test_fraction)
@@ -50,32 +44,28 @@ def run(use_deep_CNN, use_RNN, label, golden_first_sequence, golden_results):
 
 def test_shallow_CNN():
     run(use_deep_CNN=False, use_RNN=False, label='Shallow CNN',
-        golden_first_sequence='TTGAACAAGGTGAGTAATTCTAATAAGGCTGTTCAAATATGTTCCGTGTC'
-                              'AATGTTATTAACAATCAGTAGAACAGTTCCCCTTATCTTAGTTAACGTGT',
-        golden_results=OrderedDict((('Loss', 1.613392511974697),
-                                    ('Balanced accuracy', 50.0),
-                                    ('auROC', 0.581453634085213),
-                                    ('auPRC', 0.48312846202300236),
-                                    ('Recall at 5% FDR', 0.0),
-                                    ('Recall at 10% FDR', 0.0),
-                                    ('Recall at 20% FDR', 0.0),
-                                    ('Num Positives', 19),
-                                    ('Num Negatives', 21))))
+        golden_results=OrderedDict([('Loss', 0.70371496533279465),
+                                    ('Balanced accuracy', 55.639097744360896),
+                                    ('auROC', 0.50877192982456143),
+                                    ('auPRC', 0.58026674651508325),
+                                    ('Recall at 5% FDR', 9.5238095238095237),
+                                    ('Recall at 10% FDR', 9.5238095238095237),
+                                    ('Recall at 20% FDR', 9.5238095238095237),
+                                    ('Num Positives', 21),
+                                    ('Num Negatives', 19)]))
 
 
 def test_deep_CNN():
     run(use_deep_CNN=True, use_RNN=False, label='Deep CNN',
-        golden_first_sequence='AACTCTGCTGATCTATTAGAGCTACTATCGTCCAAAGCCCTCGCTACTGC'
-                              'TAGGATTATTGCTGAAGAGGAAGTAAATAATTTTTATTACCAATGCATGT',
-        golden_results=OrderedDict((('Loss', 0.9269361595503689),
-                                     ('Balanced accuracy', 50.0),
-                                     ('auROC', 0.34335839598997497),
-                                     ('auPRC', 0.4227729924796752),
-                                     ('Recall at 5% FDR', 0.0),
-                                     ('Recall at 10% FDR', 0.0),
-                                     ('Recall at 20% FDR', 0.0),
-                                     ('Num Positives', 21),
-                                     ('Num Negatives', 19))))
+        golden_results=OrderedDict([('Loss', 0.68411321005526782),
+                                    ('Balanced accuracy', 45.833333333333329),
+                                    ('auROC', 0.51822916666666663),
+                                    ('auPRC', 0.41738642611750432),
+                                    ('Recall at 5% FDR', 0.0),
+                                    ('Recall at 10% FDR', 0.0),
+                                    ('Recall at 20% FDR', 0.0),
+                                    ('Num Positives', 16),
+                                    ('Num Negatives', 24)]))
 
 
 if __name__ == '__main__':