📋 What Our Data Looks Like

# Our dataset: Breast vs Myeloid cancers
# Metadata columns + gene effect scores

model_id    cell_line_name  oncotree_lineage  oncotree_primary_disease    A1BG    A1CF    A2M
ACH-000004  HEL            Myeloid           Acute Myeloid Leukemia      0.005   -0.069  -0.098
ACH-000017  SK-BR-3        Breast            Invasive Breast Carcinoma   -0.032  -0.102  -0.013  
ACH-000019  MCF7           Breast            Invasive Breast Carcinoma   0.036   0.018   0.095
ACH-000028  KPL-1          Breast            Invasive Breast Carcinoma   -0.188  -0.149  0.077
...         ...            ...               ...                         ...     ...     ...

# We'll compare: Are certain genes more essential in breast vs myeloid cancers?
# Perfect for groupby analysis!

👀 What You'll Discover

🗺️ Our Analysis Roadmap

🧬 Filtering Our Cancer Data

# Filter for breast cancer cell lines
breast_df = df[df['oncotree_lineage'] == 'Breast']
print(f"Found {len(breast_df)} breast cancer cell lines")

# Filter for myeloid cancer cell lines
myeloid_df = df[df['oncotree_lineage'] == 'Myeloid']
print(f"Found {len(myeloid_df)} myeloid cancer cell lines")

# Advanced: Get both types in one DataFrame
both_types = df[(df['oncotree_lineage'] == 'Breast') |
                     (df['oncotree_lineage'] == 'Myeloid')]

# Alternative using .isin()
cancer_types = ['Breast', 'Myeloid']
both_types = df[df['oncotree_lineage'].isin(cancer_types)]

🧬 Our Gene Analysis Workflow

# Step 1: Filter for breast cancer (we already learned this!)
breast_df = df.loc[df['oncotree_lineage'] == 'Breast']
print(f"Found {len(breast_df)} breast cancer cell lines")

# Step 2: Select only gene effect columns
metadata_cols = ['model_id', 'cell_line_name', 'oncotree_lineage', 'oncotree_primary_disease']
gene_columns = [col for col in df.columns if col not in metadata_cols]
print(f"Analyzing {len(gene_columns)} genes")

# Step 3: Calculate mean gene effect for each gene across all breast cancer lines
breast_gene_means = breast_df[gene_columns].mean()
print("Sample of gene means:")
print(breast_gene_means.head())

# Example output:
# A1BG     -0.021
# A1CF     -0.088
# A2M      -0.013
# A2ML1     0.041
# A4GALT   -0.003

🧬 Finding Essential Genes in Our Data

# After filtering and calculating means for breast cancer
breast_df = df.loc[df['oncotree_lineage'] == 'Breast']
gene_columns = ['A1BG', 'A1CF', 'A2M', ...]  # All gene columns
breast_means = breast_df[gene_columns].mean()

# Sort to find most essential genes (most negative scores)
most_essential = breast_means.sort_values(ascending=True)
print("Top 10 most essential genes in breast cancer:")
print(most_essential.head(10))

# Or find genes that inhibit growth (most positive)
growth_inhibitors = breast_means.sort_values(ascending=False)
print("Top 10 growth inhibitor genes:")
print(growth_inhibitors.head(10))