Table 1.

Different scenarios of increasing challenge in identifying predictive biomarkers

	Fully separate pred/prog biomarkers?	Correlated biomarkers?	Interaction terms?	Subgroups?
M-1
M-2	✓
M-3	✓	✓
M-4, M-5	✓	✓	✓
M-6, M-7	✓	✓	✓	✓

Notes: Fully-separate pred/prog biomarkers is where there are no biomarkers with both predictive and prognostic strength, so a method cannot find a predictive biomarker by simply picking up on its prognostic nature. Correlated covariates creates situations where we might mistakenly pick up a noisy/prognostic biomarker, as it may be correlated to the predictive one for which we are searching. Interaction terms creates situations where two biomarkers interact to cause the outcome, which needs to be accounted for in the biomarker discovery algorithm. The presence of Subgroups creates situations where clearly defined groups of patients have enhanced treatment effect. More ticks equate to a more challenging scenarios.