#!/usr/bin/env python # -*- coding: utf-8 -*- """ Create additional modules for genes not covered by existing modules Ensure that all model genes can participate in training Run from the repository root: python model_dependencies/biomod_data_all/create_uncovered_biomods.py \ --merged_biomods model_dependencies/biomod_data_all/merged_biomods.json \ --output tutorial_runs/biomod_data/merged_modules_full.json \ --biomod_size 50 """ import json import os import csv import argparse from pathlib import Path PROJECT_ROOT = Path(__file__).resolve().parents[2] ASSET_ROOT = PROJECT_ROOT / 'model_dependencies' MODULE_DIR = Path(__file__).resolve().parent def create_uncovered_biomods( merged_modules_path, gene_list_path, gene_mapping_path, output_path, module_size=50, verbose=True ): """ Create additional modules for uncovered genes Args: merged_modules_path: merged module file path gene_list_path: model gene list path gene_mapping_path: Gene mapping file path output_path: output file path module_size: the size of the new module verbose: whether to print detailed information """ # Load model gene list with open(gene_list_path, 'r') as f: model_genes = set(line.strip() for line in f) # Load existing modules with open(merged_modules_path, 'r') as f: modules = json.load(f) # Load gene map (symbol -> ensembl) symbol_to_ens = {} ens_to_symbol = {} with open(gene_mapping_path, 'r') as f: reader = csv.DictReader(f) for row in reader: symbol = row.get('Approved symbol', '').strip() ens_id = row.get('Ensembl gene ID', '').strip() if symbol and ens_id: symbol_to_ens[symbol] = ens_id ens_to_symbol[ens_id] = symbol # Find genes that are covered covered_genes_ens = set() for module_name, genes in modules.items(): for g in genes: ens_id = symbol_to_ens.get(g) if ens_id and ens_id in model_genes: covered_genes_ens.add(ens_id) elif g in model_genes: covered_genes_ens.add(g) # Find uncovered genes uncovered_genes = model_genes - covered_genes_ens uncovered_genes = sorted(list(uncovered_genes)) if verbose: print(f"Total number of model genes: {len(model_genes)}") print(f"Genes covered: {len(covered_genes_ens)} ({len(covered_genes_ens)/len(model_genes)*100:.1f}%)") print(f"Genes not covered: {len(uncovered_genes)} ({len(uncovered_genes)/len(model_genes)*100:.1f}%)") # Create new modules for uncovered genes # Use Ensembl ID (because the model uses Ensembl ID) new_modules = {} num_new_modules = (len(uncovered_genes) + module_size - 1) // module_size for i in range(num_new_modules): start_idx = i * module_size end_idx = min((i + 1) * module_size, len(uncovered_genes)) module_genes = uncovered_genes[start_idx:end_idx] # Try converting to symbol if there is a mapping, otherwise keep the Ensembl ID module_genes_final = [] for g in module_genes: symbol = ens_to_symbol.get(g) if symbol: module_genes_final.append(symbol) else: module_genes_final.append(g) new_modules[f"uncovered_module_{i+1:04d}"] = module_genes_final if verbose: print(f"\nCreated {len(new_modules)} new modules (each approx. {module_size} genes)") # Merge existing modules and new modules merged_modules = {**modules, **new_modules} # save os.makedirs(os.path.dirname(output_path) if os.path.dirname(output_path) else '.', exist_ok=True) with open(output_path, 'w', encoding='utf-8') as f: json.dump(merged_modules, f, indent=2) if verbose: print(f"\nFinal result:") print(f" Original modules: {len(modules)}") print(f" New modules: {len(new_modules)}") print(f" Total modules: {len(merged_modules)}") print(f" Save to: {output_path}") return merged_modules def main(): parser = argparse.ArgumentParser(description='Create modules for uncovered genes') parser.add_argument('--merged_biomods', '--merged_modules', type=str, default=str(MODULE_DIR / 'merged_biomods.json'), help='Path to merged modules JSON file', dest='merged_modules') parser.add_argument('--gene_list', type=str, default=str(ASSET_ROOT / 'selected_genes_27k.txt'), help='Path to model gene list file') parser.add_argument('--gene_mapping', type=str, default=str(ASSET_ROOT / 'gene2ens.csv'), help='Path to gene mapping CSV file') parser.add_argument('--output', '-o', type=str, default=str(MODULE_DIR / 'merged_modules_full.json'), help='Output file path') parser.add_argument('--biomod_size', '--module_size', type=int, default=50, help='Size of new modules for uncovered genes', dest='module_size') args = parser.parse_args() create_uncovered_biomods( merged_modules_path=args.merged_modules, gene_list_path=args.gene_list, gene_mapping_path=args.gene_mapping, output_path=args.output, module_size=args.module_size ) if __name__ == '__main__': main()