Abstract Recent breakthrough in checkpoint blockade drugs achieved remarkable success in treating late stage tumors, although a substantial fraction of patients failed to respond. Despite growing efforts, the clinical and molecular prognostic predictors of immunotherapy response remain elusive. We developed a novel computational method to deconvolute six tumor-infiltrating immune cells from the molecular profiles of over 10,000 samples across 23 cancer types, and validated the estimates using simulations, orthogonal estimates, and pathology. Correlating the computationally-inferred immune infiltrates with patient clinical features, viral infection status, and cancer genetic alterations, we discovered associations not only extensively supported by previous studies, but also novel ones such as B cell infiltration with better outcome in glioblastoma. Analysis of cancer/testis antigen expression and CD8 T-cell abundance suggested that MAGEA3 is potentially effective in melanoma but not in NSCLC, and implicated CALR3 and SPAG5 as alternative cancer vaccine targets. We also observed PD-1 and CTLA4 expression to be associated with CD8 T-cell abundance in most cancers, supporting combined use of checkpoint blockade drugs. Finally, we found that melanomas expressing high levels of CTLA4 showed two distinct levels of CD8 T-cell infiltration, which may influence clinical responses to anti-CTLA4 agents. Taken together, our systematic analyses of tumor immunity have the potential to inform effective cancer vaccines and checkpoint blockade therapies. Citation Format: Bo Li. Comprehensive analyses of tumor immunity with implications to cancer immunotherapies. [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2015 Nov 5-9; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2015;14(12 Suppl 2):Abstract nr PR04.