Liu-Hy's picture
Add files using upload-large-folder tool
0db4514 verified
# Path Configuration
from tools.preprocess import *
# Processing context
trait = "Colon_and_Rectal_Cancer"
cohort = "GSE46517"
# Input paths
in_trait_dir = "../DATA/GEO/Colon_and_Rectal_Cancer"
in_cohort_dir = "../DATA/GEO/Colon_and_Rectal_Cancer/GSE46517"
# Output paths
out_data_file = "./output/preprocess/1/Colon_and_Rectal_Cancer/GSE46517.csv"
out_gene_data_file = "./output/preprocess/1/Colon_and_Rectal_Cancer/gene_data/GSE46517.csv"
out_clinical_data_file = "./output/preprocess/1/Colon_and_Rectal_Cancer/clinical_data/GSE46517.csv"
json_path = "./output/preprocess/1/Colon_and_Rectal_Cancer/cohort_info.json"
# STEP1
from tools.preprocess import *
# 1. Identify the paths to the SOFT file and the matrix file
soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
# 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("Sample Characteristics Dictionary:")
print(sample_characteristics_dict)
# 1. Gene Expression Data Availability
# Based on the background info: "RNA was extracted and run on the ... microarray chip"
# This indicates standard gene expression data is very likely available.
is_gene_available = True
# 2. Variable Availability and Data Type Conversion
# 2.1 Data Availability
# We look for the trait "Colon_and_Rectal_Cancer" in the sample characteristics.
# No rows show consistent data indicating colon/rectal cancer as the primary trait.
# Therefore, we consider that trait data is NOT available.
trait_row = None
# For age: row 7 contains multiple entries of "age at time of resection: ...",
# indicating distinct numeric values. Thus, age data is available at row 7.
age_row = 7
# For gender: row 8 has both "gender: male" and "gender: female", hence it
# carries at least two distinct values. So let's use row 8 for gender.
gender_row = 8
# 2.2 Data Type Conversion
def convert_trait(raw_value: str) -> int:
"""
Since trait data is not available (trait_row=None),
this function is not expected to be called.
However, we define a stub to maintain consistency.
"""
return None
def convert_age(raw_value: str) -> float:
"""
Convert age string (e.g. 'age at time of resection: 72y 4m')
to a numeric value in years (float). If parsing fails, return None.
"""
try:
# The value after the colon might look like '72y 4m'
# We'll extract that part and parse the years.
value_part = raw_value.split(':', 1)[-1].strip() # '72y 4m'
# Split on space => ['72y', '4m'] or just one piece if months missing
parts = value_part.split()
# The first part is something like '72y'
year_str = parts[0].lower().replace('y', '')
year_val = float(year_str)
return year_val
except Exception:
return None
def convert_gender(raw_value: str) -> int:
"""
Convert gender string (e.g. 'gender: male' or 'gender: female')
to a binary (female=0, male=1). If parsing fails, return None.
"""
try:
value_part = raw_value.split(':', 1)[-1].strip().lower() # 'male' or 'female'
if value_part == 'female':
return 0
elif value_part == 'male':
return 1
else:
return None
except Exception:
return None
# 3. Save Metadata
# Perform initial filtering. Trait is not available, so is_trait_available=False.
# This dataset will fail initial filtering due to missing trait, but we still log metadata.
is_trait_available = (trait_row is not None)
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=is_trait_available
)
# 4. Clinical Feature Extraction
# We only do this if trait_row is not None. Here, trait_row = None, so we skip extraction.
# End of this step.
# 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])
# The gene identifiers resemble Affymetrix probe set IDs, which are not official gene symbols.
# Therefore, these identifiers will need to be mapped to gene symbols.
print("\nrequires_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))
# STEP: Gene Identifier Mapping
# 1. From the annotation preview, we see 'ID' corresponds to the probe ids in gene_data.index,
# and 'Gene Symbol' holds the corresponding gene symbols.
mapping_df = get_gene_mapping(gene_annotation, prob_col="ID", gene_col="Gene Symbol")
# 2. Convert probe-level measurements to gene-level expression by applying the mapping.
gene_data = apply_gene_mapping(gene_data, mapping_df)
# Print a brief summary to confirm successful mapping
print("Gene-level expression data dimensions:", gene_data.shape)
# STEP7
# 1. Normalize the obtained gene data with the 'normalize_gene_symbols_in_index' function from the library.
normalized_gene_data = normalize_gene_symbols_in_index(gene_data)
normalized_gene_data.to_csv(out_gene_data_file)
# According to previous steps, we found that trait data is not available (trait_row was None),
# so is_trait_available is False.
is_trait_available = False
if not is_trait_available:
# 5. Conduct final validation to record metadata. Since we have no trait data, the dataset won't be usable.
# We must provide 'df' and 'is_biased' to the function; passing an empty DataFrame and is_biased=True
# ensures it is marked as not usable.
is_usable = validate_and_save_cohort_info(
is_final=True,
cohort=cohort,
info_path=json_path,
is_gene_available=True, # We do have gene data
is_trait_available=False, # Trait is not available
is_biased=True, # This will mark it as not usable
df=pd.DataFrame(), # Placeholder DataFrame
note="Trait data not available; dataset is not usable."
)
# Since trait is unavailable, we must skip linking or saving any final linked data.
else:
# If trait data were available, we would proceed with linking and further steps.
# But since it is not, this branch is never entered.
pass