Abstract Cancer is fundamentally a disease of disordered gene expression. In fact, reversal or neutralization of the changes in gene expression has been shown to be attractive targets for the development of new anti-cancer drugs and therapeutic strategies. New approaches such as antibody drug conjugates (ADCs) also target differentially expressed genes as a mean to recognize tumor cells to selectively deliver toxins to a tumor. In the past decade, the global analysis of gene expression in human cancers have led to the development of a number of potential new biomarkers. For instance, mesothelin (MSLN) was identified as an over-expressed gene in pancreatic cancer and later was proved to be a useful diagnostic marker and so a therapeutic target. Large-scale gene expression analysis, using techniques such as RNA sequencing, provides a powerful tool to identify genes involved in human cancers. In this study, with the ultimate goal being to identify potential novel targets for cancer immunotherapy, we conducted a pan-cancer differential expression analysis in RNA sequencing data from more than 5,000 patients with 25 different cancer types generated by The Cancer Genome Atlas (TCGA). We identified differentially expressed genes (present in at least 5% of samples in a tumor type) in comparison to a large compendium of normal transcriptomes (more than 650 samples, including 30 tissue types) gathered from Genotype-Tissue Expression (GTEx), illumina BodyMap project 2.0, TCGA, and an in-house database. In total, we identified 892 putative tumor-associated differentially expressed genes. In order to further identify novel candidate genes and rank them based on their antigenic potential, we performed an extensive literature search and systematic review to collect the characteristics of an ideal tumor antigen (TA). We developed an Analytic Hierarchy Process (AHP) model - a multiple-criteria decision-making solution - to depict antigen properties for ranking and prioritizing the tumor-associated differentially expressed genes. Our model recognizes the known tumor antigens (such as CA9, Nectin-4, FN1, MSLN and MUC16, which are currently in clinic or pre-clinical studies) in the top 25 of the ranked list. We are currently validating the top-ranked novel antigens in an orthogonal panel of tumor and normal tissues and cell lines using PCR. Note: This abstract was not presented at the meeting. Citation Format: Daryanaz Dargahi, Richard D. Swayze, Leanna Yee, Peter J. Bergqvist, Bradley J. Hedberg, Alireza Heravi-Moussavi, Edie M. Dullaghan, Ryan Dercho, Christopher Bond, Jianghong An, John S. Babcook, Steven JM Jones. Pan-cancer identification and prioritization of cancer-associated differentially expressed genes: A biomarker discovery application. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 2187. doi:10.1158/1538-7445.AM2015-2187