============================================================
OXPHOS Inhibitor Drug Validation for MSI-H
============================================================

PRISM DRUG AVAILABILITY
  Metformin:   NOT IN PRISM
  Venetoclax:  NOT IN PRISM  
  IACS-10759:  AVAILABLE (OXPHOS Complex I inhibitor)
  Tigecycline: AVAILABLE (Mitochondrial ribosome inhibitor)

DRUG SENSITIVITY (MSI-H vs MSS)
  Pan-cancer / IACS-10759:   d=-0.203, p=0.18, FDR=0.36
  Pan-cancer / tigecycline:  d=-0.206, p=0.12, FDR=0.36
  Uterus / IACS-10759:       d=-0.764, p=0.23, FDR=0.36 (underpowered, n=9)
  Bowel / tigecycline:       d=-0.284, p=0.24, FDR=0.36

VERDICT: Trend toward MSI-H sensitivity but NOT statistically significant.

CRISPR-PRISM CONCORDANCE (within MSI-H lines)
  Tigecycline vs COX5A:  r=-0.353, p=0.038 *  (n=35)
  Tigecycline vs TFAM:   r=-0.341, p=0.045 *  (n=35)
  Tigecycline vs NDUFS4: r=-0.271, p=0.115    (n=35)
  
  IACS-10759: No significant concordance within MSI-H lines

INTERPRETATION:
  - PRISM drug data provides WEAK support for OXPHOS vulnerability
  - Within MSI-H lines, tigecycline sensitivity correlates with CRISPR
    OXPHOS dependency, supporting the biological hypothesis
  - Metformin/venetoclax (clinically relevant) not testable in PRISM
  - CRISPR genetic data (33/181 mito genes, p=3.32e-08) remains
    stronger evidence than PRISM pharmacological data
