9e75f44f1a
Source-only snapshot of the cluster branch for WES execution. Large reference files (HPA/MANE/ensemble FASTA, model weights, ~597 MB) are omitted: they are baked into the container images at build time and mounted from the dreamdock-data PVC at runtime, and exceed the Gitea request size limit. Pipeline entry point is main.nf, which orchestrates the biotransformer, conplex and tissue modules as a single workflow. Ligand inputs are read from the eureka workspace; protein_zarr and chembl_db come from the dreamdock-data PVC.
146 lines
6.7 KiB
Python
Executable File
146 lines
6.7 KiB
Python
Executable File
import argparse
|
|
import pandas as pd
|
|
import os
|
|
import numpy as np
|
|
import re
|
|
import subprocess
|
|
import sys
|
|
|
|
def parse_fasta_seq(fasta_file):
|
|
try:
|
|
with open(fasta_file, 'r') as file:
|
|
return [line.strip() for line in file if not line.startswith('>')]
|
|
except Exception as e:
|
|
raise ValueError(f"Error parsing FASTA file: {fasta_file}") from e
|
|
|
|
def parse_fasta_name(fasta_file):
|
|
try:
|
|
with open(fasta_file, 'r') as file:
|
|
return [line.strip()[1:] for line in file if line.startswith('>')]
|
|
except Exception as e:
|
|
raise ValueError(f"Error parsing FASTA file: {fasta_file}") from e
|
|
|
|
def create_tsv_frags(fastas, smis, screening_batch_size):
|
|
protein_sequence_list = parse_fasta_seq(fastas)
|
|
protein_name_list = parse_fasta_name(fastas)
|
|
full_df = pd.read_csv(smis, delimiter='\t', names=['smiles', 'chem_id'])
|
|
|
|
#splits = int(np.ceil((len(protein_sequence_list))/screening_batch_size))
|
|
#protein_sequence_list = np.array_split(protein_sequence_list, splits)
|
|
#protein_name_list = np.array_split(protein_name_list, splits)
|
|
|
|
try:
|
|
[full_df.assign(protein_id = [protein_id]*full_df.shape[0]).assign(protein_sequence = [protein_sequence]*full_df.shape[0])[['protein_id', 'chem_id', 'protein_sequence', 'smiles']].to_csv(f"{protein_id}.tsv", sep='\t', header=False, index=False) for protein_id, protein_sequence in zip(protein_name_list,protein_sequence_list)]
|
|
tsv_files = [f"{protein_id}.tsv" for protein_id in protein_name_list]
|
|
return tsv_files
|
|
|
|
except Exception as e:
|
|
print(f"Failed to process files: {e}")
|
|
raise
|
|
|
|
|
|
def screen_frags(tsv_files):
|
|
#[subprocess.run(["conplex-dti", "predict","--data-file", tsv_file,"--model-path", "/home/omic/ConPLex/models/ConPLex_v1_BindingDB.pt","--outfile", f"{os.path.splitext(os.path.basename(tsv_file))[0]}_results.tsv"]) for tsv_file in tsv_files]
|
|
[subprocess.run(["conplex-dti", "predict","--data-file", tsv_file,"--model-path", "/home/omic/ConPLex/models/Run_best_model_epoch46.pt","--outfile", f"{os.path.splitext(os.path.basename(tsv_file))[0]}_results.tsv"]) for tsv_file in tsv_files]
|
|
scores_files = [f"{os.path.splitext(os.path.basename(tsv_file))[0]}_results.tsv" for tsv_file in tsv_files]
|
|
#scores_files = []
|
|
#for tsv_file in tsv_files:
|
|
#base_name = os.path.splitext(os.path.basename(tsv_file))[0]
|
|
#scores_file = f"{base_name}_scores.tsv"
|
|
|
|
#subprocess.run([
|
|
# "conplex-dti", "predict",
|
|
# "--data-file", tsv_file,
|
|
# "--model-path", "/home/omic/ConPLex/models/ConPLex_v1_BindingDB.pt",
|
|
# "--outfile", f"{base_name}_results.tsv"
|
|
#])
|
|
|
|
#with open(scores_file, 'w') as f:
|
|
# f.write("chem_id\tprotein_id\tscore\n")
|
|
|
|
#with open(f"{base_name}_results.tsv", 'r') as results_file, open(f"{base_name}_temp.tsv", 'w') as temp_file:
|
|
# for line in results_file:
|
|
# fields = line.strip().split('\t')
|
|
# temp_file.write('\t'.join(fields) + '\n')
|
|
|
|
#merged_df = pd.read_csv(f"{base_name}_temp.tsv", sep='\t', names=['chem_id', 'protein_id', 'score'])
|
|
#tsv_df = pd.read_csv(tsv_file, sep='\t', names=['protein_id', 'chem_id', 'sequence', 'smiles'])
|
|
|
|
#merged_df = merged_df.merge(tsv_df[['chem_id', 'smiles']], on='chem_id')
|
|
#merged_df.to_csv(scores_file, sep='\t', index=False)
|
|
|
|
#os.remove(f"{base_name}_temp.tsv")
|
|
#scores_files.append(scores_file)
|
|
return scores_files
|
|
|
|
def collect_frags(scores_files, fasta_file, outdir, split_write_df = 'no'):
|
|
def concatenate_dataframes(file_list):
|
|
df_list = []
|
|
header = None
|
|
for i, file in enumerate(file_list):
|
|
if os.path.getsize(file) > 0:
|
|
if i == 0:
|
|
df = pd.read_csv(file, sep='\t', index_col=None)
|
|
if df.columns[0] == 'chem_id':
|
|
header = df.columns
|
|
else:
|
|
df = pd.read_csv(file, sep='\t', index_col=None, header=None)
|
|
df.columns = ['chem_id', 'protein_id', 'score', 'smiles']
|
|
else:
|
|
df = pd.read_csv(file, sep='\t', index_col=None, header=None)
|
|
if header is not None:
|
|
df.columns = header
|
|
df_list.append(df)
|
|
else:
|
|
print(f"Skipping empty file: {file}")
|
|
if df_list:
|
|
concatenated_df = pd.concat(df_list, ignore_index=True)
|
|
return concatenated_df
|
|
else:
|
|
return pd.DataFrame()
|
|
|
|
scores_df = concatenate_dataframes(scores_files)
|
|
|
|
if not scores_df.empty:
|
|
score_column_index = 2 if 'score' not in scores_df.columns else 'score'
|
|
scores_df[score_column_index] = pd.to_numeric(scores_df[score_column_index], errors='coerce')
|
|
scores_df = scores_df.dropna(subset=[score_column_index])
|
|
sorted_scores_df = scores_df.sort_values(by=score_column_index, ascending=False)
|
|
if 'score' not in scores_df.columns:
|
|
sorted_scores_df.columns = ['chem_id', 'protein_id', 'score', 'hit', 'smiles']
|
|
|
|
base_name = os.path.splitext(os.path.basename(fasta_file))[0]
|
|
all_scores_file = os.path.join(outdir, f"{base_name}_all_scores.tsv")
|
|
sorted_scores_df.to_csv(all_scores_file, sep='\t', index=False, header=True)
|
|
|
|
else:
|
|
print("No score files found or all score files are empty.")
|
|
|
|
with open(fasta_file, 'r') as fasta:
|
|
fasta_lines = fasta.readlines()
|
|
fasta_sequence = fasta_lines[1].strip() if len(fasta_lines) > 1 else ''
|
|
|
|
if not scores_df.empty:
|
|
df = sorted_scores_df
|
|
header = df.columns.tolist()
|
|
|
|
if split_write_df == 'yes':
|
|
for index, row in df.iterrows():
|
|
row_df = pd.DataFrame([row], columns=header)
|
|
csv_filename = f"{row['complex_name'].replace(' ', '-')}_hit.csv"
|
|
row_df.to_csv(os.path.join(outdir, csv_filename), index=False)
|
|
|
|
if __name__ == "__main__":
|
|
parser = argparse.ArgumentParser(description="Process and screen fragments.")
|
|
parser.add_argument("--fastas", required=True, help="Path to the FASTA file.")
|
|
parser.add_argument("--smis", required=True, help="Path to the SMILES file.")
|
|
parser.add_argument("--screening_batch_size", type=int, default=100000, help="Batch size for screening.")
|
|
parser.add_argument("--outdir", default="output", help="Output directory.")
|
|
parser.add_argument("--split_write_df", default="no", help="Write all complexes in separate files")
|
|
|
|
args = parser.parse_args()
|
|
|
|
tsv_files = create_tsv_frags(args.fastas, args.smis, args.screening_batch_size)
|
|
scores_files = screen_frags(tsv_files)
|
|
#collect_frags(scores_files, args.fastas, args.outdir, args.split_write_df)
|