-
Notifications
You must be signed in to change notification settings - Fork 2
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
Showing
5 changed files
with
123 additions
and
79 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -1,64 +1,102 @@ | ||
| #!/usr/bin/env python | ||
| # TODO: rewrite this | ||
| # | ||
| # import argparse | ||
| # | ||
| # import h5py | ||
| # import keras.backend as K | ||
| # from keras.models import model_from_json | ||
| # | ||
| # from evolutron.engine import Model as DeepTrainer | ||
| # from evolutron.extra_layers import custom_layers | ||
| # from evolutron.motifs import motif_extraction | ||
| # from evolutron.tools import Handle, load_dataset | ||
| # | ||
| # | ||
| # # Check if package is installed, else fallback to developer mode imports | ||
| # | ||
| # | ||
| # def main(filename, data_id): | ||
| # # First load model architecture | ||
| # hf = h5py.File(filename) | ||
| # model_config = hf.attrs['model_config'].decode('utf8') | ||
| # hf.close() | ||
| # | ||
| # net = DeepTrainer(model_from_json(model_config, custom_objects=custom_layers)) | ||
| # | ||
| # # Then load model parameters | ||
| # net.load_all_param_values(filename) | ||
| # | ||
| # handle = Handle.from_filename(filename) | ||
| # | ||
| # if data_id == 'model': | ||
| # data_id = handle.dataset | ||
| # | ||
| # x_data, y_data = load_dataset(data_id, padded=True, max_aa=net.input._keras_shape[1]) | ||
| # | ||
| # conv_layers = net.get_conv_layers() | ||
| # | ||
| # for depth, conv_layer in enumerate(conv_layers): | ||
| # conv_scores = conv_layer.output | ||
| # # Compile function that spits out the outputs of the correct convolutional layer | ||
| # boolean_mask = K.any(K.not_equal(net.input, 0.0), axis=-1, keepdims=True) | ||
| # conv_scores = conv_scores * K.cast(boolean_mask, K.floatx()) | ||
| # | ||
| # custom_fun = K.function([net.input], [conv_scores]) | ||
| # # Start visualizations | ||
| # motif_extraction(custom_fun, x_data, conv_layer.filters, | ||
| # conv_layer.kernel_size[0], handle, depth) | ||
| # | ||
| # | ||
| # if __name__ == '__main__': | ||
| # parser = argparse.ArgumentParser(description='Network visualization module.') | ||
| # parser.add_argument("model", help='Path to the file') | ||
| # | ||
| # parser.add_argument("-d", "--dataset", type=str, default='model', | ||
| # help='Dataset on which the motifs will be generated upon. Write "model" to infer' \ | ||
| # 'automatically from model.') | ||
| # | ||
| # args = parser.parse_args() | ||
| # | ||
| # kwargs = {'filename': args.model, | ||
| # 'data_id': args.dataset} | ||
| # | ||
| # main(**kwargs) | ||
| # coding=utf-8 | ||
| import math | ||
| import os | ||
| import sys | ||
|
|
||
| import keras.backend as K | ||
| import pandas as pd | ||
| from absl import flags | ||
| from keras.utils import Sequence | ||
|
|
||
| from evolutron.engine import load_model | ||
| from evolutron.extra_layers import custom_layers | ||
| from evolutron.motifs import motif_extraction | ||
| from evolutron.tools import preprocess_dataset | ||
|
|
||
| flags.DEFINE_string('infile', '', 'The input file. THe supported format currently is a TSV output from UniProt.', ) | ||
| flags.DEFINE_string('model_file', '', 'The model file for the model to use for predictions.') | ||
| flags.DEFINE_string("output_dir", os.path.expandvars(os.curdir), 'The directory to store CoMET output data.') | ||
| flags.DEFINE_enum("motifs_filetype", 'txt+png', ['png', 'pdf', 'txt', 'txt+pdf', 'txt+png'], | ||
| 'Choose between different file types to save the extracted motifs from CoMET.' | ||
| 'A typical workflow for subsequent analysis would be to extract the motifs as text files (txt) and' | ||
| 'then use the tool sites2meme to transform them to MEME format and submit for search in MAST.') | ||
|
|
||
| FLAGS = flags.FLAGS | ||
|
|
||
| try: | ||
| FLAGS(sys.argv) | ||
| except flags.Error as e: | ||
| print(e) | ||
| print(FLAGS) | ||
| sys.exit(1) | ||
|
|
||
|
|
||
| class UniProtSequence(Sequence): | ||
|
|
||
| def __init__(self, dataframe, batch_size, max_dim): | ||
| self.x = dataframe | ||
| self.batch_size = batch_size | ||
| self.max_dim = max_dim | ||
|
|
||
| def __len__(self): | ||
| return math.ceil(len(self.x) / self.batch_size) | ||
|
|
||
| def __getitem__(self, idx): | ||
| batch_x = self.x.iloc[idx * self.batch_size:(idx + 1) * self.batch_size] | ||
| return self.process_batch(batch_x) | ||
|
|
||
| def process_batch(self, batch): | ||
| return preprocess_dataset(batch.Sequence, padded=True, max_aa=self.max_dim, min_aa=self.max_dim) | ||
|
|
||
|
|
||
| def extract_motifs(model, proteins, handle): | ||
| # TODO: refactor motif extraction | ||
| # data_gen = UniProtSequence(proteins, 100, model.input_shape[1]) | ||
|
|
||
| for depth, conv_layer in enumerate(model.get_conv_layers()): | ||
| conv_scores = conv_layer.output | ||
| # Compile function that spits out the outputs of the correct convolutional layer | ||
| boolean_mask = K.any(K.not_equal(model.input, 0.0), axis=-1, keepdims=True) | ||
| conv_scores = conv_scores * K.cast(boolean_mask, K.floatx()) | ||
|
|
||
| # motif_extractor = Model([model.input], [conv_scores]) | ||
| custom_fun = K.function([model.input], [conv_scores]) | ||
|
|
||
| x_data = preprocess_dataset(proteins.Sequence, padded=True, max_aa=model.input_shape[1], | ||
| min_aa=model.input_shape[1]) | ||
|
|
||
| # Start motif extraction | ||
| motif_extraction(custom_fun, x_data, conv_layer.filters, conv_layer.kernel_size[0], handle, depth, | ||
| filetype=FLAGS.motifs_filetype) | ||
|
|
||
|
|
||
| def main(): | ||
| try: | ||
| model = load_model(FLAGS.model_file, custom_objects=custom_layers) | ||
| except Exception as e: | ||
| print('Unable to load model.') | ||
| print(e) | ||
| sys.exit(1) | ||
|
|
||
| model_key = FLAGS.model_file.split('/')[-1].split('.')[0] | ||
|
|
||
| try: | ||
| proteins = pd.read_csv(FLAGS.infile, sep='\t') | ||
| except Exception as e: | ||
| print('Unable to read input protein dataset.') | ||
| print(e) | ||
| sys.exit(1) | ||
|
|
||
| dataset_key = FLAGS.infile.split('/')[-1].split('.')[0] | ||
|
|
||
| output_folder = os.path.join(FLAGS.output_dir, 'motifs', dataset_key, model_key) | ||
|
|
||
| if not os.path.isdir(output_folder): | ||
| os.makedirs(output_folder) | ||
|
|
||
| extract_motifs(model, proteins, output_folder) | ||
|
|
||
|
|
||
| if __name__ == "__main__": | ||
| main() |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters