A transposable element atlas of aggressive B‐cell non‐Hodgkin lymphomas defines novel classifications of Burkitt lymphoma independent of EBV status

Introduction: Clinical and pathological heterogeneity in aggressive B-cell non-Hodgkin lymphomas may be driven by several mechanisms of oncogenesis, such as translocations involving the immunoglobulin locus proto-oncogenes, ‘cell-of-origin’ from germinal center (GC), viral associations. Previous classifications have been defined one or more these characteristics transforming events. However, transposable elements (TEs), which make up roughly half human genome key regulatory functions, previously overlooked both healthy GC lymphomas. We hypothesized that unique TE signatures, part endogenous retroviruses (HERVs), would allow a novel categorization Here we present comprehensive, locus-specific atlas expression B cells, diffuse large lymphoma (DLBCL), Epstein-Barr Virus (EBV)-positive negative Burkitt (BL), follicular (FL). Methods: obtained RNA-seq data 529 DLBCLs belonging to TCGA NCICCR cohorts, 113 sporadic endemic BL, 12 FL CGCI cohort for gene quantification. sorted cells were two recent studies (Holmes et al., 2020; Agirre 2019). TEs quantified with Telescope, addresses computational challenges repetitive interspersed reads, low read counts using an expectation-maximization algorithm. used STAR indexing alignment, HTSeq quantifying EBV reads BL samples, DESeq2 identify differentially expressed TEs, LRT, Boruta, Lasso feature selection. Results: TE-driven clustering showed HERVs could independently distinguish stages cell differentiation specific subsets, plasmablasts bone marrow plasma having highest number HERVs. found HERV loci interest upstream POU5F1B MYC can differentiate other cells. Furthermore, describe map DLBCL, FL. Strikingly, find subdivided into three HERV-driven clusters, are not gene-only clustering, 4 sufficient between clusters. Clusters independent status, clinical variant, location translocation, display high LINE elements, lncRNAs, snoRNAs. Conclusions: Here, report FL, provide evidence disease-specific changes selectively subtypes opportunities subclassification biological relevance. The research was funded by: work supported US National Institutes Health (NIH) grant CA260691 (DFN), UM1AI164559 (DFN). MLB is Department Medicine, Fund Future program at Weill Cornell Medicine sponsored Elsa Miller Foundation. J.L.M. Medical Scientist Training Program Cornell–Rockefeller–Sloan Kettering Tri-Institutional MD-PhD (T32GM007739). Keywords: Aggressive lymphoma, Bioinformatics, Computational Systems Biology, Pathology Classification Lymphomas No conflicts interests pertinent abstract.