p3/preprocess/Osteoarthritis/code/GSE107105.py

# Path Configuration
from tools.preprocess import *

# Processing context
trait = "Osteoarthritis"
cohort = "GSE107105"

# Input paths
in_trait_dir = "../DATA/GEO/Osteoarthritis"
in_cohort_dir = "../DATA/GEO/Osteoarthritis/GSE107105"

# Output paths
out_data_file = "./output/preprocess/3/Osteoarthritis/GSE107105.csv"
out_gene_data_file = "./output/preprocess/3/Osteoarthritis/gene_data/GSE107105.csv"
out_clinical_data_file = "./output/preprocess/3/Osteoarthritis/clinical_data/GSE107105.csv"
json_path = "./output/preprocess/3/Osteoarthritis/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
# Given it's a microarray analysis of gene expression, gene data is available
is_gene_available = True

# 2.1 Data Availability
# Trait is available in row 0 (disease)
trait_row = 0 
# Age is available in row 1
age_row = 1
# Gender is available in row 2 (Sex)
gender_row = 2

# 2.2 Data Type Conversion
def convert_trait(value):
    """Convert disease status to binary (1 for OA, 0 for RA)"""
    if not value:
        return None
    value = value.split(': ')[-1].strip()
    if value == 'OA':
        return 1
    elif value == 'RA':
        return 0
    return None

def convert_age(value):
    """Convert age to continuous numeric value"""
    if not value:
        return None
    try:
        return float(value.split(': ')[-1].strip())
    except:
        return None

def convert_gender(value):
    """Convert gender to binary (0 for Female, 1 for Male)"""
    if not value:
        return None
    value = value.split(': ')[-1].strip()
    if value.lower() == 'female':
        return 0
    elif value.lower() == 'male':
        return 1
    return None

# 3. Save Metadata
is_trait_available = trait_row is not None
_ = validate_and_save_cohort_info(is_final=False, 
                                cohort=cohort,
                                info_path=json_path,
                                is_gene_available=is_gene_available,
                                is_trait_available=is_trait_available)

# 4. Clinical Feature Extraction
if trait_row is not None:
    selected_clinical = geo_select_clinical_features(
        clinical_df=clinical_data,
        trait=trait,
        trait_row=trait_row,
        convert_trait=convert_trait,
        age_row=age_row,
        convert_age=convert_age,
        gender_row=gender_row,
        convert_gender=convert_gender
    )
    
    # Preview the processed data
    preview = preview_df(selected_clinical)
    print("Preview of processed clinical data:")
    print(preview)
    
    # Save to CSV
    selected_clinical.to_csv(out_clinical_data_file)
# 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])
# The gene identifiers are numeric (16650001, etc.) which appear to be probe IDs
# from a microarray platform. These need to be mapped to human gene symbols.

requires_gene_mapping = True
# Extract gene annotation data from platform table
# First try the simple approach to see what columns we get
gene_annotation = get_gene_annotation(soft_file_path)

print("Shape:", gene_annotation.shape)

print("\nFirst few rows:")
print(gene_annotation.head())

print("\nColumn names:")
print(list(gene_annotation.columns))

print("\nUnique values in selected columns:")
for col in gene_annotation.columns:
    uniq_vals = gene_annotation[col].drop_duplicates().head()
    print(f"\n{col}:")
    print(uniq_vals.tolist())
# Get gene annotation data with both markers for table boundaries
gene_annotation = get_gene_annotation(soft_file_path)

print("Gene annotation columns:")
print(gene_annotation.columns)

print("\nFirst few rows:")
print(gene_annotation.head())

# Create mapping dataframe with probe IDs and gene symbols
probe_gene_map = pd.DataFrame()
probe_gene_map['ID'] = gene_annotation['ID_REF'].astype(str) 
probe_gene_map['Gene'] = gene_annotation['GENE'].fillna('')

# Apply the mapping to convert probe data to gene data
gene_data = apply_gene_mapping(genetic_data, probe_gene_map)

print("\nGene mapping dataframe shape:", probe_gene_map.shape)
print("\nFirst few rows of gene mapping:")
print(probe_gene_map.head())

print("\nGene expression dataframe shape:", gene_data.shape)
print("\nFirst few rows of gene expression data:")
print(gene_data.head())

# Save the gene expression data
gene_data.to_csv(out_gene_data_file)
# Extract gene annotation data from platform table using default prefixes
gene_annotation = get_gene_annotation(soft_file_path)

# Display structure and content
print("Data shape:", gene_annotation.shape)
print("\nPreview of first few rows:")
print(gene_annotation.head(3)) 

# If needed, try additional filtering to ensure we get the gene mapping table
gene_annotation = gene_annotation[gene_annotation['ID'].notna()].copy()

# Verify we have proper ID column matching our expression data
print("\nFirst few IDs:")
print(gene_annotation['ID'].head())

# Check for potential gene symbol columns
potential_symbol_cols = []
for col in gene_annotation.columns:
    sample_vals = gene_annotation[col].dropna().astype(str).head()
    # Check if values look like gene symbols (capital letters, numbers)
    if any(val.isupper() and len(val) < 10 for val in sample_vals):
        potential_symbol_cols.append(col)
        print(f"\nPotential gene symbol column '{col}' values:")
        print(sample_vals)
# 1. Cannot proceed with gene normalization since gene mapping failed in previous steps
# This dataset may not have gene symbols in its annotation
print("WARNING: Gene symbols are unavailable in the annotation data. Proceeding with probe-level analysis.")

# Use probe IDs as features
probe_data = genetic_data

# Reload clinical data that was processed earlier
selected_clinical_df = pd.read_csv(out_clinical_data_file, index_col=0)

# 2. Link clinical and probe-level data 
linked_data = pd.concat([selected_clinical_df, probe_data], axis=0).T

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

# 4. Check for bias in trait and demographic features
trait_biased, linked_data = judge_and_remove_biased_features(linked_data, trait)

# 5. Final validation and information saving
note = "Gene expression data from synovial fibroblasts, comparing osteoarthritis (OA) vs rheumatoid arthritis (RA). Analysis uses probe IDs since gene symbol mapping was unavailable."
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
)

# 6. Save linked data only if usable
if is_usable:
    os.makedirs(os.path.dirname(out_data_file), exist_ok=True)
    # Save probe-level data
    linked_data.to_csv(out_data_file)
    
    # Also save probe expression data separately
    probe_data.to_csv(out_gene_data_file)