p3/preprocess/Esophageal_Cancer/code/TCGA.py

# Path Configuration
from tools.preprocess import *

# Processing context
trait = "Esophageal_Cancer"

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

# Output paths
out_data_file = "./output/preprocess/3/Esophageal_Cancer/TCGA.csv"
out_gene_data_file = "./output/preprocess/3/Esophageal_Cancer/gene_data/TCGA.csv"
out_clinical_data_file = "./output/preprocess/3/Esophageal_Cancer/clinical_data/TCGA.csv"
json_path = "./output/preprocess/3/Esophageal_Cancer/cohort_info.json"

# Select the ESCA (Esophageal Cancer) cohort directory
cohort_dir = os.path.join(tcga_root_dir, 'TCGA_Esophageal_Cancer_(ESCA)')

# Get paths for clinical and genetic data files  
clinical_file, genetic_file = tcga_get_relevant_filepaths(cohort_dir)

# Load the data files
clinical_df = pd.read_csv(clinical_file, index_col=0, sep='\t')
genetic_df = pd.read_csv(genetic_file, index_col=0, sep='\t')

# Print clinical data columns for review
print("Clinical data columns:")
print(clinical_df.columns.tolist())
# 1. Identify candidate columns for age and gender
candidate_age_cols = ['age_at_initial_pathologic_diagnosis', 'age_began_smoking_in_years', 'days_to_birth']
candidate_gender_cols = ['gender']

# 2. Preview the data
# First check available directories
print("Available directories in TCGA root:")
print(os.listdir(tcga_root_dir))

# Get clinical data path using actual directory structure
cohort_dir = os.path.join(tcga_root_dir, "TCGA-ESCA") 
clinical_file_path, _ = tcga_get_relevant_filepaths(cohort_dir)

# Read clinical data 
clinical_df = pd.read_csv(clinical_file_path, index_col=0)

# Extract candidate columns
age_data = clinical_df[candidate_age_cols]
gender_data = clinical_df[candidate_gender_cols]

# Preview data as dictionaries
print("\nAge columns preview:")
print(preview_df(age_data))
print("\nGender columns preview:") 
print(preview_df(gender_data))
# Select the ESCA (Esophageal Cancer) cohort directory
cohort_dir = os.path.join(tcga_root_dir, 'TCGA_Esophageal_Cancer_(ESCA)')

# Get paths for clinical and genetic data files  
clinical_file, genetic_file = tcga_get_relevant_filepaths(cohort_dir)

# Load the data files
clinical_df = pd.read_csv(clinical_file, index_col=0, sep='\t')
genetic_df = pd.read_csv(genetic_file, index_col=0, sep='\t')

# Print clinical data columns for review
print("Clinical data columns:")
print(clinical_df.columns.tolist())
# Check values in candidate columns 
age_col = "age_at_initial_pathologic_diagnosis"

# Gender column is straightforward
gender_col = "gender"

# Print chosen columns
print(f"Selected age column: {age_col}")
print(f"Selected gender column: {gender_col}")
# Carry over the selected demographic columns
age_col = "age_at_initial_pathologic_diagnosis"
gender_col = "gender"

# 1. Extract and standardize clinical features 
clinical_df = tcga_select_clinical_features(clinical_df, trait, age_col, gender_col)

# 2. Normalize gene expression data
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="TCGA_Esophageal_Cancer_(ESCA)",
    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="Data from TCGA Esophageal Cancer 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)