Created
April 11, 2022 14:34
-
-
Save leiterenato/031d64357b39b055047a21b4eae1b016 to your computer and use it in GitHub Desktop.
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
# Copyright 2021 Google LLC | |
# | |
# Licensed under the Apache License, Version 2.0 (the "License"); | |
# you may not use this file except in compliance with the License. | |
# You may obtain a copy of the License at | |
# | |
# http://www.apache.org/licenses/LICENSE-2.0 | |
# | |
# Unless required by applicable law or agreed to in writing, software | |
# distributed under the License is distributed on an "AS IS" BASIS, | |
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | |
# See the License for the specific language governing permissions and | |
# limitations under the License. | |
from typing import Mapping, List | |
from kfp.v2 import dsl | |
from src import config | |
from src.components.jackhmmer import jackhmmer as JackhmmerOp | |
from src.components.hhblits import hhblits as HhblitsOp | |
from src.components.hhsearch import hhsearch as HhsearchOp | |
from src.components.aggregate_features import aggregate_features as AggregateOp | |
from src.components.model_predict import predict as PredictOp | |
from src.components.relax_protein import relax as RelaxOp | |
@dsl.pipeline( | |
name=config.PIPELINE_NAME, | |
description=config.PIPELINE_DESCRIPTION | |
) | |
def alphafold_pipeline( | |
sequence_path: str, | |
project: str='alphafold-dev', | |
region: str='us-central1', | |
max_template_date: str='2020-05-14', | |
models: List[Mapping]=[{'model_name': 'model_1', 'random_seed': 1}], | |
use_gpu_for_relaxation: bool=True, | |
num_ensemble: int=1, | |
reference_datasets_uri: str=config.REFERENCE_DATASETS_URI, | |
model_params_uri: str=config.MODEL_PARAMS_GCS_LOCATION | |
): | |
"""Runs AlphaFold inference.""" | |
input_sequence = dsl.importer( | |
artifact_uri=sequence_path, | |
artifact_class=dsl.Dataset, | |
reimport=True) | |
input_sequence.set_display_name('Input sequence') | |
model_parameters = dsl.importer( | |
artifact_uri=model_params_uri, | |
artifact_class=dsl.Artifact, | |
reimport=True) | |
model_parameters.set_display_name('Model parameters') | |
reference_databases = dsl.importer( | |
artifact_uri=reference_datasets_uri, | |
artifact_class=dsl.Dataset, | |
reimport=False, | |
metadata={ | |
config.UNIREF90: config.UNIREF90_PATH, | |
config.MGNIFY: config.MGNIFY_PATH, | |
config.BFD: config.BFD_PATH, | |
config.UNICLUST30: config.UNICLUST30_PATH, | |
config.PDB70: config.PDB70_PATH, | |
config.PDB_MMCIF: config.PDB_MMCIF_PATH, | |
config.PDB_OBSOLETE: config.PDB_OBSOLETE_PATH, | |
config.PDB_SEQRES: config.PDB_SEQRES_PATH, | |
config.UNIPROT: config.UNIPROT_PATH, | |
} | |
) | |
reference_databases.set_display_name('Reference databases') | |
search_uniref = JackhmmerOp( | |
project=project, | |
region=region, | |
database=config.UNIREF90, | |
reference_databases=reference_databases.output, | |
sequence=input_sequence.output, | |
components_image=config.ALPHAFOLD_COMPONENTS_IMAGE | |
) | |
search_uniref.set_display_name('Search Uniref')#.set_caching_options(enable_caching=True) | |
search_mgnify = JackhmmerOp( | |
project=project, | |
region=region, | |
database=config.MGNIFY, | |
reference_databases=reference_databases.output, | |
sequence=input_sequence.output, | |
components_image=config.ALPHAFOLD_COMPONENTS_IMAGE | |
) | |
search_mgnify.set_display_name('Search Mgnify')#.set_caching_options(enable_caching=True) | |
search_uniclust = HhblitsOp( | |
project=project, | |
region=region, | |
msa_dbs=[config.UNICLUST30], | |
reference_databases=reference_databases.output, | |
sequence=input_sequence.output, | |
components_image=config.ALPHAFOLD_COMPONENTS_IMAGE | |
) | |
search_uniclust.set_display_name('Search Uniclust')#.set_caching_options(enable_caching=True) | |
search_bfd = HhblitsOp( | |
project=project, | |
region=region, | |
msa_dbs=[config.BFD], | |
reference_databases=reference_databases.output, | |
sequence=input_sequence.output, | |
components_image=config.ALPHAFOLD_COMPONENTS_IMAGE | |
) | |
search_bfd.set_display_name('Search BFD')#.set_caching_options(enable_caching=True) | |
search_pdb = HhsearchOp( | |
project=project, | |
region=region, | |
template_dbs=[config.PDB70], | |
mmcif_db=config.PDB_MMCIF, | |
obsolete_db=config.PDB_OBSOLETE, | |
max_template_date=max_template_date, | |
reference_databases=reference_databases.output, | |
sequence=input_sequence.output, | |
msa=search_uniref.outputs['msa'], | |
components_image=config.ALPHAFOLD_COMPONENTS_IMAGE | |
) | |
search_pdb.set_display_name('Search Pdb')#.set_caching_options(enable_caching=True) | |
aggregate_features = AggregateOp( | |
sequence=input_sequence.output, | |
msa1=search_uniref.outputs['msa'], | |
msa2=search_mgnify.outputs['msa'], | |
msa3=search_bfd.outputs['msa'], | |
msa4=search_uniclust.outputs['msa'], | |
template_features=search_pdb.outputs['template_features'], | |
) | |
aggregate_features.set_display_name('Aggregate features')#.set_caching_options(enable_caching=True) | |
# Think what to do with random seed when switch to Parallel loop | |
with dsl.ParallelFor(models) as model: | |
model_predict = PredictOp( | |
model_features=aggregate_features.outputs['features'], | |
model_params=model_parameters.output, | |
model_name=model.model_name, | |
num_ensemble=num_ensemble, | |
random_seed=model.random_seed | |
) | |
model_predict.set_display_name('Predict')#.set_caching_options(enable_caching=True) | |
model_predict.set_cpu_limit(config.CPU_LIMIT) | |
model_predict.set_memory_limit(config.MEMORY_LIMIT) | |
model_predict.set_gpu_limit(config.GPU_LIMIT) | |
model_predict.add_node_selector_constraint(config.GKE_ACCELERATOR_KEY, config.GPU_TYPE) | |
model_predict.set_env_variable("TF_FORCE_UNIFIED_MEMORY", config.TF_FORCE_UNIFIED_MEMORY) | |
model_predict.set_env_variable("XLA_PYTHON_CLIENT_MEM_FRACTION", config.XLA_PYTHON_CLIENT_MEM_FRACTION) | |
relax_protein = RelaxOp( | |
unrelaxed_protein=model_predict.outputs['unrelaxed_protein'], | |
use_gpu=use_gpu_for_relaxation, | |
) | |
relax_protein.set_display_name('Relax protein')#.set_caching_options(enable_caching=True) | |
relax_protein.set_cpu_limit(config.RELAX_CPU_LIMIT) | |
relax_protein.set_memory_limit(config.RELAX_MEMORY_LIMIT) | |
relax_protein.set_gpu_limit(config.RELAX_GPU_LIMIT) | |
relax_protein.add_node_selector_constraint(config.GKE_ACCELERATOR_KEY, config.RELAX_GPU_TYPE) | |
relax_protein.set_env_variable("TF_FORCE_UNIFIED_MEMORY", config.TF_FORCE_UNIFIED_MEMORY) | |
relax_protein.set_env_variable("XLA_PYTHON_CLIENT_MEM_FRACTION", config.XLA_PYTHON_CLIENT_MEM_FRACTION) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment