output/preprocess/Essential_Thrombocythemia/code/TCGA.py

# Path Configuration
from tools.preprocess import *

# Processing context
trait = "Essential_Thrombocythemia"

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

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


# Step 1: Initial Data Loading
import os
import pandas as pd

# Step 1: Review subdirectories and select the best match for Essential Thrombocythemia (ET)
provided_subdirs = [
    'TCGA_lower_grade_glioma_and_glioblastoma_(GBMLGG)', 'TCGA_Uterine_Carcinosarcoma_(UCS)',
    'TCGA_Thyroid_Cancer_(THCA)', 'TCGA_Thymoma_(THYM)', 'TCGA_Testicular_Cancer_(TGCT)',
    'TCGA_Stomach_Cancer_(STAD)', 'TCGA_Sarcoma_(SARC)', 'TCGA_Rectal_Cancer_(READ)',
    'TCGA_Prostate_Cancer_(PRAD)', 'TCGA_Pheochromocytoma_Paraganglioma_(PCPG)',
    'TCGA_Pancreatic_Cancer_(PAAD)', 'TCGA_Ovarian_Cancer_(OV)', 'TCGA_Ocular_melanomas_(UVM)',
    'TCGA_Mesothelioma_(MESO)', 'TCGA_Melanoma_(SKCM)', 'TCGA_Lung_Squamous_Cell_Carcinoma_(LUSC)',
    'TCGA_Lung_Cancer_(LUNG)', 'TCGA_Lung_Adenocarcinoma_(LUAD)', 'TCGA_Lower_Grade_Glioma_(LGG)',
    'TCGA_Liver_Cancer_(LIHC)', 'TCGA_Large_Bcell_Lymphoma_(DLBC)', 'TCGA_Kidney_Papillary_Cell_Carcinoma_(KIRP)',
    'TCGA_Kidney_Clear_Cell_Carcinoma_(KIRC)', 'TCGA_Kidney_Chromophobe_(KICH)',
    'TCGA_Head_and_Neck_Cancer_(HNSC)', 'TCGA_Glioblastoma_(GBM)', 'TCGA_Esophageal_Cancer_(ESCA)',
    'TCGA_Endometrioid_Cancer_(UCEC)', 'TCGA_Colon_and_Rectal_Cancer_(COADREAD)',
    'TCGA_Colon_Cancer_(COAD)', 'TCGA_Cervical_Cancer_(CESC)', 'TCGA_Breast_Cancer_(BRCA)',
    'TCGA_Bladder_Cancer_(BLCA)', 'TCGA_Bile_Duct_Cancer_(CHOL)', 'TCGA_Adrenocortical_Cancer_(ACC)',
    'TCGA_Acute_Myeloid_Leukemia_(LAML)'
]

# Define trait-related keywords for ET; avoid broad terms that could incorrectly match unrelated cohorts
keywords = ['essential thrombocythemia', 'thrombocythemia', 'myeloproliferative', 'mpn', 'polycythemia', 'myelofibrosis']

selected_subdirs = [d for d in provided_subdirs if any(k in d.lower() for k in keywords)]
selected_subdir = selected_subdirs[0] if selected_subdirs else None

clinical_df = None
genetic_df = None

if selected_subdir is None:
    # No suitable TCGA cohort for Essential Thrombocythemia; record skip and finish
    _ = validate_and_save_cohort_info(
        is_final=False,
        cohort="TCGA",
        info_path=json_path,
        is_gene_available=False,
        is_trait_available=False
    )
    print("No suitable TCGA cohort found for Essential Thrombocythemia. Skipping.")
else:
    # Step 2: Identify clinical and genetic file paths
    cohort_dir = os.path.join(tcga_root_dir, selected_subdir)
    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)
    genetic_df = pd.read_csv(genetic_file_path, sep='\t', index_col=0, low_memory=False)

    # Step 4: Print clinical columns
    print(clinical_df.columns.tolist())