p3/preprocess/Mesothelioma/code/GSE163722.py

# Path Configuration
from tools.preprocess import *

# Processing context
trait = "Mesothelioma"
cohort = "GSE163722"

# Input paths
in_trait_dir = "../DATA/GEO/Mesothelioma"
in_cohort_dir = "../DATA/GEO/Mesothelioma/GSE163722"

# Output paths
out_data_file = "./output/preprocess/3/Mesothelioma/GSE163722.csv"
out_gene_data_file = "./output/preprocess/3/Mesothelioma/gene_data/GSE163722.csv"
out_clinical_data_file = "./output/preprocess/3/Mesothelioma/clinical_data/GSE163722.csv"
json_path = "./output/preprocess/3/Mesothelioma/cohort_info.json"

# Get file paths
soft_file_path, matrix_file_path = geo_get_relevant_filepaths(in_cohort_dir)

# Get background info and clinical data
background_info, clinical_data = get_background_and_clinical_data(matrix_file_path)
print("Background Information:")
print(background_info)
print("\nSample Characteristics:")

# Get dictionary of unique values per row 
unique_values_dict = get_unique_values_by_row(clinical_data)
for row, values in unique_values_dict.items():
    print(f"\n{row}:")
    print(values)
# 1. Gene Expression Data Availability
# Study title indicates RERG expression analysis
is_gene_available = True

# 2. Variable Availability and Data Type Conversion
# 2.1 Data Availability
# All samples are tumor samples (constant trait), so trait data is not useful
trait_row = None   
age_row = None    # No age data 
gender_row = None # No gender data

# 2.2 Data Type Conversion Functions
def convert_trait(x):
    if pd.isna(x):
        return None
    value = x.split(": ")[-1].strip().lower()
    # Since trait data is constant, this function won't be used
    return None

def convert_age(x):
    # Not needed since age data not available
    return None

def convert_gender(x):
    # Not needed since gender data not available
    return None

# 3. Save Metadata
# is_trait_available is False since trait values are constant
validate_and_save_cohort_info(is_final=False, 
                            cohort=cohort,
                            info_path=json_path,
                            is_gene_available=is_gene_available,
                            is_trait_available=False)

# 4. Clinical Feature Extraction
# Skip since trait_row is None (constant trait values)
# Get gene expression data from matrix file
genetic_data = get_genetic_data(matrix_file_path)

# Examine data structure
print("Data structure and head:")
print(genetic_data.head())

print("\nShape:", genetic_data.shape)

print("\nFirst 20 row IDs (gene/probe identifiers):")
print(list(genetic_data.index)[:20])

# Get a few column names to verify sample IDs
print("\nFirst 5 column names:")
print(list(genetic_data.columns)[:5])
# Reviewing gene identifiers: they appear to be simple row numbers
# This suggests they are probe IDs rather than gene symbols
# We need to map them to actual gene symbols
requires_gene_mapping = True
# Extract gene annotation data
gene_annotation = get_gene_annotation(soft_file_path)

# Display column names and preview data
print("Column names:")
print(gene_annotation.columns)

print("\nPreview of gene annotation data:")
print(preview_df(gene_annotation))
# Looking at the gene identifiers, we need a different method to match them to gene data 
# Let's modify the annotation dataframe to add position indices
gene_annotation['position'] = gene_annotation.index.astype(str)

# Extract gene identifier column (position) and gene symbol column, with proper renaming
mapping_data = gene_annotation[['position', 'gene_assignment']]
mapping_data = mapping_data.rename(columns={'position': 'ID', 'gene_assignment': 'Gene'})

# Apply the mapping to get gene-level expression data
gene_data = apply_gene_mapping(genetic_data, mapping_data)

# Normalize gene symbols by checking against a dictionary of gene synonyms
gene_data = normalize_gene_symbols_in_index(gene_data)

# Save the processed gene data
gene_data.to_csv(out_gene_data_file)

# Print the shape and preview the result
print("\nFinal gene data shape:", gene_data.shape)
print("\nPreview of final gene data:")
print(preview_df(gene_data))
# Create minimal linked data with constant trait
linked_data = gene_data.T
linked_data['Mesothelioma'] = 1  # all samples are tumor
linked_data = handle_missing_values(linked_data, 'Mesothelioma')

# Check for bias
trait_biased, linked_data = judge_and_remove_biased_features(linked_data, 'Mesothelioma')

# Final validation
note = "Dataset contains only tumor samples without control samples for comparison."
is_usable = validate_and_save_cohort_info(
    is_final=True,
    cohort=cohort, 
    info_path=json_path,
    is_gene_available=True,
    is_trait_available=True,
    is_biased=trait_biased,
    df=linked_data,
    note=note
)

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