The current era that allows multiple forms of genomic data to be collected offers great opportunity and challenge to effectively integrate these data with clinical information for gaining biological insights that can guide the development of more effective therapies. We recently developed the CC-PROMISE procedure, which combines canonical correlation (CC) analysis and projection onto the most interesting statistical evidence (PROMISE) to integrate two forms of molecular data with multiple pharmacologic and clinical outcomes. When applied to the analysis of a pediatric AML data set, the CC-PROMISE procedure identified patterns of association among DNA methylation, RNA expression, disease response, and prognosis, which were statistically significant, biologically meaningful, and clinically relevant. The finding that the greater genome-wide methylation burden is associated with a greater risk of relapse provides scientific rationale for using demethylating agents in the ongoing clinical trial AML16 (clinicaltrials.gov ID: NCT03164057).
Cao X, Crews KR, Downing J, Lamba J, Pounds SB. CC-PROMISE effectively integrates two forms of molecular data with multiple biologically related endpoints. BMC Bioinformatics, 17: 382, 2016.