Abstract Background: Tissue samples are routinely preserved as FFPE blocks. This abundant sample reservoir holds promises in clinical omics for disease diagnosis and precision cancer medicine. However, due to extensive cross-linking, nucleic acids are not efficiently extracted from FFPE samples and are heavily degraded, posing challenges for molecular assays including next generation sequencing. Our study aims to identify the RNA sequencing (RNA-seq) library prep method for FFPE samples that provides comparable data to the established poly-A enrichment method for FF tissue. Methods: RNA-seq data were collected on matched FF and FFPE samples from triple negative breast cancer (TNBC; n=7) and normal breast tissues (n=2). Tissues were sourced from a commercial vendor. Three library preparation protocols from Illumina were tested: poly(A) enrichment protocol (Poly-A; FF only), ribosomal RNA deletion protocol (Ribo-Zero Gold), and exome-capture based protocol (RNA-Access). Data were compared on absolute and differential gene expression, as well as molecular subtyping for breast cancer (PAM50; Parker 2009). Results: Both absolute expression values and log fold changes between tumor and normal samples correlate strongly across different tissue preservation and library preparation methods. Among tested methods, RNA-Access had the highest yield of mappable reads on coding region and provided uniform transcript coverage with less rRNA contamination. Regardless of the protocol, all TNBC samples were classified as PAM50 basal-like subtype. As expected, FF with either Ribo-Zero or RNA-Access had the strongest correlation with FF PolyA on the distance to subtype centroids (r=0.99). For FFPE methods, RNA-Access (r=0.94) had the strongest correlation with FF PolyA, followed by Ribo-Zero (r=0.85). Principal component analysis showed that samples were primarily segregated by technical differences (library preparation or tissue preservation method) rather than biological differences (tumor/normal status or patient identification), suggesting strong technical bias in the data. Gene families and biological processes underlying the differentially expressed genes due to technical variations included ribosomal, histone and metabolic pathways proteins. The aforementioned technical bias can be largely eliminated by a batch effect adjustment method (COMBAT; Johnson 2007). Conclusion: As expected, RNA-seq data is primarily impacted by technical differences. RNA-Access based transcriptome sequencing on breast FFPE samples generated highly concordant gene expression profiles when compared to data derived from FF samples. Our study supports the use of RNA-Access based transcriptome sequencing on FFPE samples when FF samples are not available. Citation Format: Xi Zhao, Marianna Zavodovskaya, Luting Zhuo, Kevin Kwei, Xin Guo, Zhaoshi Jiang, Scott D. Patterson, Carrie B. Brachmann. Systematic evaluation of transcriptome sequencing shows comparable profiles for an exome-capture method for formalin-fixed, paraffin-embedded (FFPE) breast cancer tissues and the standard poly-A method for matched fresh frozen (FF) tissues [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 3833. doi:10.1158/1538-7445.AM2017-3833