Liu-Hy's picture
Add files using upload-large-folder tool
61e25af verified
# Path Configuration
from tools.preprocess import *
# Processing context
trait = "Obesity"
cohort = "GSE123088"
# Input paths
in_trait_dir = "../DATA/GEO/Obesity"
in_cohort_dir = "../DATA/GEO/Obesity/GSE123088"
# Output paths
out_data_file = "./output/preprocess/3/Obesity/GSE123088.csv"
out_gene_data_file = "./output/preprocess/3/Obesity/gene_data/GSE123088.csv"
out_clinical_data_file = "./output/preprocess/3/Obesity/clinical_data/GSE123088.csv"
json_path = "./output/preprocess/3/Obesity/cohort_info.json"
# Get file paths
soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
# Extract background info and clinical data
background_info, clinical_data = get_background_and_clinical_data(matrix_file)
# Get unique values per clinical feature
sample_characteristics = get_unique_values_by_row(clinical_data)
# Print background info
print("Dataset Background Information:")
print(f"{background_info}\n")
# Print sample characteristics
print("Sample Characteristics:")
for feature, values in sample_characteristics.items():
print(f"Feature: {feature}")
print(f"Values: {values}\n")
import pandas as pd
import numpy as np
# Check gene expression data availability
is_gene_available = True # CD4+ T cells expression data should be present
# Find trait, age and gender data rows and define conversion functions
trait_row = 1 # 'primary diagnosis' contains obesity info
age_row = 3 # 'age' info starts in row 3, continues in row 4
gender_row = 2 # 'Sex' information
def convert_trait(x):
if pd.isna(x):
return None
val = x.split(': ')[1] if ': ' in x else x
if val.upper() in ['OBESITY']:
return 1
elif 'CONTROL' in val.upper():
return 0
return None
def convert_age(x):
if pd.isna(x):
return None
try:
val = x.split(': ')[1] if ': ' in x else x
return float(val)
except:
return None
def convert_gender(x):
if pd.isna(x):
return None
val = x.split(': ')[1] if ': ' in x else x
if val.upper() == 'FEMALE':
return 0
elif val.upper() == 'MALE':
return 1
return None
# Validate and save initial cohort info
_ = validate_and_save_cohort_info(
is_final=False,
cohort=cohort,
info_path=json_path,
is_gene_available=is_gene_available,
is_trait_available=trait_row is not None
)
# Extract clinical features if trait data is available
if trait_row is not None:
selected_clinical_df = geo_select_clinical_features(
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 clinical features
preview = preview_df(selected_clinical_df)
print("Clinical features preview:")
print(preview)
# Save clinical features
selected_clinical_df.to_csv(out_clinical_data_file)
# Extract gene expression data from matrix file
gene_data = get_genetic_data(matrix_file)
# Print first 20 row IDs and shape of data to help debug
print("Shape of gene expression data:", gene_data.shape)
print("\nFirst few rows of data:")
print(gene_data.head())
print("\nFirst 20 gene/probe identifiers:")
print(gene_data.index[:20])
# Inspect a snippet of raw file to verify identifier format
import gzip
with gzip.open(matrix_file, 'rt', encoding='utf-8') as f:
lines = []
for i, line in enumerate(f):
if "!series_matrix_table_begin" in line:
# Get the next 5 lines after the marker
for _ in range(5):
lines.append(next(f).strip())
break
print("\nFirst few lines after matrix marker in raw file:")
for line in lines:
print(line)
# Based on the gene expression data shown, the identifiers appear to be numerical indices (1, 2, 3, etc.)
# rather than human gene symbols. This indicates mapping to gene symbols will be required.
requires_gene_mapping = True
# Get file paths using library function
soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
# First inspect a snippet of raw SOFT file to understand its structure
import gzip
print("Inspecting raw SOFT file structure:")
with gzip.open(soft_file, 'rt', encoding='utf-8') as f:
lines = []
for i, line in enumerate(f):
if i < 20: # Look at first 20 lines
lines.append(line.strip())
else:
break
print('\n'.join(lines))
print("\n" + "="*50 + "\n")
# Extract gene annotation from SOFT file, excluding header lines
gene_annotation = get_gene_annotation(soft_file)
# Preview gene annotation data
print("Gene annotation shape:", gene_annotation.shape)
print("\nGene annotation columns:")
print(gene_annotation.columns.tolist())
print("\nGene annotation preview:")
print(preview_df(gene_annotation))
print("\nNumber of non-null values in each column:")
print(gene_annotation.count())
# Print first few rows of annotation data to verify structure
print("\nFirst 5 rows of annotation data:")
print(gene_annotation.head().to_string())
# First find the SubSeries ID from the SOFT file
import gzip
subseries_id = None
with gzip.open(soft_file, 'rt', encoding='utf-8') as f:
for line in f:
if line.startswith('!Series_relation'):
if 'SubSeries of:' not in line:
subseries_id = line.strip().split(' = ')[1].split(' ')[0]
break
# The correct subseries directory should be one level up
platform_soft_file = os.path.join(os.path.dirname(os.path.dirname(soft_file)),
subseries_id,
f"{subseries_id}_family.soft.gz")
# Extract platform annotation from the subseries SOFT file
platform_lines = []
with gzip.open(platform_soft_file, 'rt', encoding='utf-8') as f:
reading_platform = False
for line in f:
if line.startswith('!Platform_table_begin'):
reading_platform = True
# Skip the header line
header = next(f).strip().split('\t')
continue
elif line.startswith('!Platform_table_end'):
break
elif reading_platform:
platform_lines.append(line.strip())
# Create platform annotation dataframe
platform_data = [line.split('\t') for line in platform_lines]
df_platform = pd.DataFrame(platform_data, columns=header)
# Get mapping using ID and gene symbol columns
mapping_data = get_gene_mapping(df_platform, 'ID', 'Gene Symbol')
# Apply gene mapping to convert probe-level data to gene expression data
gene_data = apply_gene_mapping(gene_data, mapping_data)
# Preview results
print("Shape of gene expression data after mapping:", gene_data.shape)
print("\nFirst few rows of mapped gene expression data:")
print(gene_data.head())
# Ensure index is string type for gene symbol mapping
gene_data.index = gene_data.index.astype(str)
# Convert Entrez IDs to gene symbols using the built-in mapping
gene_data = normalize_gene_symbols_in_index(gene_data)
# Print shape and preview results
print("Shape of gene expression data after mapping:", gene_data.shape)
print("\nFirst few rows of mapped gene expression data:")
print(gene_data.head())
# Save gene expression data
gene_data.to_csv(out_gene_data_file)
# 1. First get platform annotation with gene symbols
import gzip
with gzip.open(soft_file, 'rt') as f:
platform_section = False
platform_lines = []
for line in f:
if line.startswith('^PLATFORM'):
platform_section = True
elif platform_section and line.startswith('!Platform_table_begin'):
header = next(f).strip().split('\t')
for l in f:
if l.startswith('!Platform_table_end'):
break
platform_lines.append(l.strip())
# Create platform annotation dataframe
platform_data = [line.split('\t') for line in platform_lines]
platform_df = pd.DataFrame(platform_data, columns=header)
mapping_df = get_gene_mapping(platform_df, 'ID', 'Gene Symbol')
# Apply mapping to convert probe-level data to gene expression data
gene_data = apply_gene_mapping(gene_data, mapping_df)
# 2. Load clinical data and normalize gene symbols
selected_clinical = pd.read_csv(out_clinical_data_file, index_col=0)
gene_data = normalize_gene_symbols_in_index(gene_data)
os.makedirs(os.path.dirname(out_gene_data_file), exist_ok=True)
gene_data.to_csv(out_gene_data_file)
# 3. Link clinical and genetic data
linked_data = geo_link_clinical_genetic_data(selected_clinical, gene_data)
# 4. Handle missing values
linked_data = handle_missing_values(linked_data, trait)
# 5. Check for biased features and remove them if needed
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=True,
is_trait_available=True,
is_biased=is_biased,
df=linked_data,
note="Study examining gene expression in CD4+ T cells across multiple diseases including obesity"
)
# 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)
# 1. Since we have Entrez IDs in the gene expression data index,
# we can directly normalize them to gene symbols
gene_data = normalize_gene_symbols_in_index(gene_data)
os.makedirs(os.path.dirname(out_gene_data_file), exist_ok=True)
gene_data.to_csv(out_gene_data_file)
# 2. Load clinical data and link with genetic data
selected_clinical = pd.read_csv(out_clinical_data_file, index_col=0)
linked_data = geo_link_clinical_genetic_data(selected_clinical, gene_data)
# 3. Handle missing values
linked_data = handle_missing_values(linked_data, trait)
# 4. Check for biased features and remove them if needed
is_biased, linked_data = judge_and_remove_biased_features(linked_data, trait)
# 5. Validate and save cohort info
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=is_biased,
df=linked_data,
note="Study examining gene expression in CD4+ T cells across multiple diseases including obesity"
)
# 6. 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)