output/preprocess/Epilepsy/code/TCGA.py

# Path Configuration
from tools.preprocess import *

# Processing context
trait = "Epilepsy"

# Input paths
tcga_root_dir = "../DATA/TCGA"

# Output paths
out_data_file = "./output/z3/preprocess/Epilepsy/TCGA.csv"
out_gene_data_file = "./output/z3/preprocess/Epilepsy/gene_data/TCGA.csv"
out_clinical_data_file = "./output/z3/preprocess/Epilepsy/clinical_data/TCGA.csv"
json_path = "./output/z3/preprocess/Epilepsy/cohort_info.json"


# Step 1: Initial Data Loading
import os

# Step 1: Select the most relevant TCGA cohort directory for the trait "Epilepsy"
synonyms = ["epilepsy", "seizure", "seizures", "ictal", "epileptic"]
all_dirs = [d for d in os.listdir(tcga_root_dir) if os.path.isdir(os.path.join(tcga_root_dir, d))]
matched_dirs = [d for d in all_dirs if any(s in d.lower() for s in synonyms)]

selected_cohort_dirname = None
if matched_dirs:
    # Prefer exact 'epilepsy' match if present; otherwise take the first matched
    prioritized = sorted(matched_dirs, key=lambda d: (0 if "epilepsy" in d.lower() else 1, d.lower()))
    selected_cohort_dirname = prioritized[0]

if selected_cohort_dirname is None:
    print("No suitable TCGA cohort found for the trait 'Epilepsy'. Skipping this trait.")
    # Record unusable dataset status
    _ = validate_and_save_cohort_info(
        is_final=False,
        cohort="TCGA",
        info_path=json_path,
        is_gene_available=False,
        is_trait_available=False
    )
    clinical_df = None
    gene_df = None
else:
    # Step 2: Identify clinical and genetic data file paths
    cohort_dir = os.path.join(tcga_root_dir, selected_cohort_dirname)
    clinical_file_path, genetic_file_path = tcga_get_relevant_filepaths(cohort_dir)

    # Step 3: Load both files as DataFrames
    clinical_df = pd.read_csv(clinical_file_path, sep='\t', index_col=0, low_memory=False, compression='infer')
    gene_df = pd.read_csv(genetic_file_path, sep='\t', index_col=0, low_memory=False, compression='infer')

    # Step 4: Print column names of the clinical data
    print(list(clinical_df.columns))