# build CNN model
model = Sequential()
model.add(Conv2D(50, kernel_size=(20, 12), activation='relu'))
model.add(MaxPooling2D(pool_size=(1, 2)))
model.add(Flatten())
model.add(Dropout(0.3))
model.add(Dense(650, activation='relu'))
model.add(Dropout(0.3))
model.add(Dense(2, activation='softmax'))
model.compile(loss=keras.losses.categorical_crossentropy,
optimizer=keras.optimizers.Adadelta(lr=0.01),
metrics=['accuracy'])
print('Loading training data...')
pos_Train = readFile("./data/pos_training_dataset.txt", maxlen)
neg_Train = readFile("./data/neg_training_dataset_2.txt", maxlen)
print('Generating labels and features...')
(y_train, x_train) = createTrainTestData(pos_Train, neg_Train, "Onehot")
print('Shuffling the data...')
index = np.arange(len(y_train))
np.random.shuffle(index)
x_train = x_train[index, :]
y_train = y_train[index]
x_train = x_train.reshape(x_train.shape[0], seq_rows, seq_cols, 1)
y_train = keras.utils.to_categorical(y_train, num_classes)
print('Training...')
DeepT3_1 = model_from_json(json_string)
DeepT3_1.load_weights(
os.path.join(args.DeepT3_directory, 'models_weights/DeepT3_1.h5'))
with open(
os.path.join(args.DeepT3_directory,
'models_weights/DeepT3_2.json'), 'r') as f:
json_string = f.read()
DeepT3_2 = model_from_json(json_string)
DeepT3_2.load_weights(
os.path.join(args.DeepT3_directory, 'models_weights/DeepT3_2.h5'))
print('Loading data...')
testData = readFile(args.fasta_file, maxlen)
print('Generating features...')
x_test = createData(testData, "Onehot")
x_test = x_test.reshape(x_test.shape[0], seq_rows, seq_cols, 1)
print('Predicting...')
predicted_Probability_1 = DeepT3_1.predict(x_test)
predicted_Probability_2 = DeepT3_2.predict(x_test)
prediction_1 = DeepT3_1.predict_classes(x_test)
prediction_2 = DeepT3_2.predict_classes(x_test)
prediction = prediction_1 + prediction_2
faa_accessions = [i.name for i in SeqIO.parse(args.fasta_file, "fasta")]
print('Saving the result..')
# build CNN model
model = Sequential()
model.add(Conv2D(50, kernel_size=(20,14),activation='relu'))
model.add(MaxPooling2D(pool_size=(1,2)))
model.add(Flatten())
model.add(Dropout(0.3))
model.add(Dense(650, activation='relu'))
model.add(Dropout(0.3))
model.add(Dense(2, activation='softmax'))
model.compile(loss=keras.losses.categorical_crossentropy,optimizer=keras.optimizers.Adadelta(lr=0.006),metrics=['accuracy'])
print('Loading training data...')
pos_Train = readFile("./data/pos_train.txt",maxlen)
neg_Train = readFile("./data/neg_train.txt",maxlen)
print('Generating labels and features...')
(y_train, x_train) = createTrainTestData(pos_Train,neg_Train,"Onehot")
print('Shuffling the data...')
index = np.arange(len(y_train))
np.random.shuffle(index)
x_train = x_train[index,:]
y_train = y_train[index]
x_train = x_train.reshape(x_train.shape[0],seq_rows, seq_cols,1)
y_train = keras.utils.to_categorical(y_train, num_classes)
print('Training...')
print('Loading model...')
with open('./models_weights/DeepT3_1.json', 'r') as f:
json_string = f.read()
DeepT3_1 = model_from_json(json_string)
DeepT3_1.load_weights('./models_weights/DeepT3_1.h5')
with open('./models_weights/DeepT3_2.json', 'r') as f:
json_string = f.read()
DeepT3_2 = model_from_json(json_string)
DeepT3_2.load_weights('./models_weights/DeepT3_2.h5')
print('Loading data...')
testData = readFile("./data/t4sptrain263.txt", maxlen)
print('Generating features...')
x_test = createData(testData, "Onehot")
x_test = x_test.reshape(x_test.shape[0], seq_rows, seq_cols, 1)
print('Predicting...')
predicted_Probability_1 = DeepT3_1.predict(x_test)
predicted_Probability_2 = DeepT3_2.predict(x_test)
prediction_1 = DeepT3_1.predict_classes(x_test)
prediction_2 = DeepT3_2.predict_classes(x_test)
prediction = prediction_1 + prediction_2
print('Saving the result..')
f = open("result.txt", "w")
for i in prediction:
import numpy as np
maxlen = 150
seq_rows, seq_cols = 20, maxlen
print('Loading model...')
with open('./models_weights/json_model.json', 'r') as f:
json_string = f.read()
model = model_from_json(json_string)
model.load_weights('./models_weights/model_weights.h5')
print('Loading data...')
testData = readFile("./data/S.cerevisiae-RBP354.txt", maxlen)
print('Generating features...')
x_test = createData(testData, "Onehot")
x_test = x_test.reshape(x_test.shape[0], seq_rows, seq_cols, 1)
print('Predicting...')
predicted_Probability = model.predict(x_test)
prediction = model.predict_classes(x_test)
print('Saving the result..')
f = open("result.txt", "w")
for i in prediction:
if i == 1:
f.write("T4SE\n")