Liu-Hy's picture
Add files using upload-large-folder tool
a5a8278 verified
# Path Configuration
from tools.preprocess import *
# Processing context
trait = "Bone_Density"
cohort = "GSE56816"
# Input paths
in_trait_dir = "../DATA/GEO/Bone_Density"
in_cohort_dir = "../DATA/GEO/Bone_Density/GSE56816"
# Output paths
out_data_file = "./output/preprocess/1/Bone_Density/GSE56816.csv"
out_gene_data_file = "./output/preprocess/1/Bone_Density/gene_data/GSE56816.csv"
out_clinical_data_file = "./output/preprocess/1/Bone_Density/clinical_data/GSE56816.csv"
json_path = "./output/preprocess/1/Bone_Density/cohort_info.json"
# STEP1
from tools.preprocess import *
# 1. Attempt to identify the paths to the SOFT file and the matrix file
try:
soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
except AssertionError:
print("[WARNING] Could not find the expected '.soft' or '.matrix' files in the directory.")
soft_file, matrix_file = None, None
if soft_file is None or matrix_file is None:
print("[ERROR] Required GEO files are missing. Please check file names in the cohort directory.")
else:
# 2. Read the matrix file to obtain background information and sample characteristics data
background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
background_info, clinical_data = get_background_and_clinical_data(matrix_file,
background_prefixes,
clinical_prefixes)
# 3. Obtain the sample characteristics dictionary from the clinical dataframe
sample_characteristics_dict = get_unique_values_by_row(clinical_data)
# 4. Explicitly print out all the background information and the sample characteristics dictionary
print("Background Information:")
print(background_info)
print("\nSample Characteristics Dictionary:")
print(sample_characteristics_dict)
# Step 1: Determine if gene expression data is available
is_gene_available = True # Based on the series title indicating it is a gene expression study
# Step 2: Identify data availability and define conversion functions
# Observed keys in sample characteristics:
# 0 -> ['gender: Female'] (only one unique value => not useful for analysis)
# 1 -> ['bone mineral density: high BMD', 'bone mineral density: low BMD']
# 2 -> ['state: postmenopausal', 'state: premenopausal']
# 3 -> ['cell type: monocytes']
# The trait "Bone_Density" is mapped to key=1 with two unique values => binary variable
trait_row = 1
# No explicit age information => not available
age_row = None
# Gender is constant ("Female") => not available
gender_row = None
# Define the data type conversion for the trait "Bone_Density" (binary: low=0, high=1)
def convert_trait(value: str):
parts = value.split(':', 1)
if len(parts) < 2:
return None
raw_val = parts[1].strip().lower()
if 'low' in raw_val:
return 0
elif 'high' in raw_val:
return 1
return None
# Since age and gender are not available, no conversion functions are defined for them
convert_age = None
convert_gender = None
# Step 3: Conduct initial filtering and save metadata
is_usable = 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),
)
# Step 4: Clinical feature extraction 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 extracted clinical features
print(preview_df(selected_clinical_df, 5))
# Save the clinical dataframe to CSV
selected_clinical_df.to_csv(out_clinical_data_file, index=False)
# STEP3
# 1. Use the get_genetic_data function from the library to get the gene_data from the matrix_file previously defined.
gene_data = get_genetic_data(matrix_file)
# 2. Print the first 20 row IDs (gene or probe identifiers) for future observation.
print(gene_data.index[:20])
# After examining the identifiers like '1007_s_at' and others, they are Affymetrix probe set IDs.
# Therefore, they are not standard human gene symbols and require mapping to gene symbols.
print("requires_gene_mapping = True")
# STEP5
# 1. Use the 'get_gene_annotation' function from the library to get gene annotation data from the SOFT file.
gene_annotation = get_gene_annotation(soft_file)
# 2. Use the 'preview_df' function from the library to preview the data and print out the results.
print("Gene annotation preview:")
print(preview_df(gene_annotation))
# Gene Identifier Mapping
# 1. Decide which columns are the probe ID and gene symbol
# From the annotation preview, "ID" matches the probe IDs in gene_data, and "Gene Symbol" contains gene symbols.
mapping_df = get_gene_mapping(gene_annotation, prob_col='ID', gene_col='Gene Symbol')
# 2. Convert probe-level measurements to gene-level measurements
gene_data = apply_gene_mapping(gene_data, mapping_df)
# (Optional) Preview the first few rows of the mapped gene_data
print(gene_data.head())
import os
import pandas as pd
# STEP7: Data Normalization and Linking
# 1) Normalize the gene symbols in the previously obtained gene_data
normalized_gene_data = normalize_gene_symbols_in_index(gene_data)
normalized_gene_data.to_csv(out_gene_data_file)
# 2) Load clinical data only if it exists and is non-empty
if os.path.exists(out_clinical_data_file) and os.path.getsize(out_clinical_data_file) > 0:
# Read the file
clinical_temp = pd.read_csv(out_clinical_data_file)
# Adjust row index to label the trait, age, and gender properly
if clinical_temp.shape[0] == 3:
clinical_temp.index = [trait, "Age", "Gender"]
elif clinical_temp.shape[0] == 2:
clinical_temp.index = [trait, "Gender"]
elif clinical_temp.shape[0] == 1:
clinical_temp.index = [trait]
# 2) Link the clinical and normalized genetic data
linked_data = geo_link_clinical_genetic_data(clinical_temp, normalized_gene_data)
# 3) Handle missing values
linked_data = handle_missing_values(linked_data, trait)
# 4) Check for severe bias in the trait; remove biased demographic features if present
trait_biased, linked_data = judge_and_remove_biased_features(linked_data, trait)
# 5) Final quality validation and save metadata
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=f"Final check on {cohort} with {trait}."
)
# 6) If the linked data is usable, save it
if is_usable:
linked_data.to_csv(out_data_file)
else:
# If no valid clinical data file is found, finalize metadata indicating trait unavailability
is_usable = validate_and_save_cohort_info(
is_final=True,
cohort=cohort,
info_path=json_path,
is_gene_available=True,
is_trait_available=False,
is_biased=True, # Force a fallback so that it's flagged as unusable
df=pd.DataFrame(),
note=f"No trait data found for {cohort}, final metadata recorded."
)
# Per instructions, do not save a final linked data file when trait data is absent.