Transcriptional deconvolution reveals consistent functional subtypes of pancreatic cancer epithelium and stroma

Jing He,H. Carlo Maurer,Sam R. Holmstrom,Tao Su,Aqeel Ahmed,Hanina Hibshoosh,John A. Chabot,Paul E. Oberstein,Antonia R. Sepulveda,Jeanine M. Genkinger,Jiapeng Zhang,Alina C. Iuga,Andrea Califano,Mukesh Bansal,Kenneth P. Olive
DOI: https://doi.org/10.1101/288779
2018-03-26
Abstract:Summary Bulk tumor tissues comprise intermixed populations of neoplastic cells and multiple lineages of stromal cells. We used laser capture microdissection and RNA sequencing to disentangle the transcriptional programs active in the malignant epithelium and stroma of pancreatic ductal adenocarcinoma (PDA). This led to the development of a new algorithm (ADVOCATE) that accurately predicts the compartment fractions of bulk tumor samples and can computationally purify bulk gene expression data from PDA. We also present novel stromal subtypes, derived from 110 microdissected PDA stroma samples, that were enriched in extracellular matrix– and immune–associated processes. Finally, we applied ADVOCATE to systematically evaluate cross–compartment subtypes spanning four patient cohorts, revealing consistent functional classes and survival associations despite substantial compositional differences.
What problem does this paper attempt to address?