p3/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/preprocess/3/Epilepsy/TCGA.csv"
out_gene_data_file = "./output/preprocess/3/Epilepsy/gene_data/TCGA.csv"
out_clinical_data_file = "./output/preprocess/3/Epilepsy/clinical_data/TCGA.csv"
json_path = "./output/preprocess/3/Epilepsy/cohort_info.json"

# List available cohorts first
cohorts = os.listdir(tcga_root_dir)

# Try each cohort until we find one with the data files we need
clinical_df = None
for cohort in cohorts:
    try:
        cohort_dir = os.path.join(tcga_root_dir, cohort)
        clinical_file_path, _ = tcga_get_relevant_filepaths(cohort_dir)
        clinical_df = pd.read_csv(clinical_file_path, sep='\t', index_col=0)
        break
    except:
        continue

if clinical_df is not None:
    columns = clinical_df.columns.tolist()
    candidate_age_cols = [col for col in columns if 'age' in col.lower() or 'years' in col.lower()]
    candidate_gender_cols = [col for col in columns if 'gender' in col.lower() or 'sex' in col.lower()]
else:
    candidate_age_cols = []
    candidate_gender_cols = []

print(f"candidate_age_cols = {candidate_age_cols}")
print(f"candidate_gender_cols = {candidate_gender_cols}")
# Set age_col using the most suitable column from the candidate_age_cols
age_col = 'age_at_initial_pathologic_diagnosis'  # This is the standard age column in TCGA data

# Set gender_col based on candidate column(s)
gender_col = 'gender' if candidate_gender_cols else None

# Print chosen columns
print(f'Selected age column: {age_col}')
print(f'Selected gender column: {gender_col}')
# 1. Extract and standardize clinical features
clinical_df = tcga_select_clinical_features(clinical_df, trait, age_col, gender_col)

# 2. Normalize gene expression data
gene_expression_dir = os.path.join(tcga_root_dir, cohort)
_, gene_file_path = tcga_get_relevant_filepaths(gene_expression_dir)
genetic_df = pd.read_csv(gene_file_path, sep='\t', index_col=0)

normalized_genetic_df = normalize_gene_symbols_in_index(genetic_df)
os.makedirs(os.path.dirname(out_gene_data_file), exist_ok=True)
normalized_genetic_df.to_csv(out_gene_data_file)

# 3. Link clinical and genetic data
linked_data = pd.merge(normalized_genetic_df.T, clinical_df, left_index=True, right_index=True)

# Add trait labels based on sample IDs
linked_data[trait] = linked_data.index.map(tcga_convert_trait)

# 4. Handle missing values
linked_data = handle_missing_values(linked_data, trait)

# 5. Check for bias in features
is_biased, linked_data = judge_and_remove_biased_features(linked_data, trait)

# 6. Validate and save cohort info
is_usable = validate_and_save_cohort_info(
    is_final=True,
    cohort=cohort,
    info_path=json_path,
    is_gene_available=len(normalized_genetic_df.columns) > 0,
    is_trait_available=trait in linked_data.columns,
    is_biased=is_biased,
    df=linked_data,
    note=f"Data from {cohort} cohort"
)

# 7. Save linked data if usable
if is_usable:
    os.makedirs(os.path.dirname(out_data_file), exist_ok=True)
    linked_data.to_csv(out_data_file)