The presently disclosed subject matter relates to the use of one or more biomarkers to evaluate the likelihood that an EZH2 inhibitor would produce an anti-cancer effect in a subject. It is based, at least in part, on the discovery that loss of BAP1 results in the upregulation of EZH2 expression and activity. In a specific non-limiting embodiment, the method comprises obtaining a sample of the cancer from a subject, and determining, in the sample, the expression level of an BAP1 biomarker, where if the BAP1 biomarker is absent or expressed at lower level in the cancer as compared to a reference control level, then administering a therapeutically effective amount of an EZH2 inhibitor to produce an anti-cancer effect.
