- Main
Big data discovery of cancer immunotherapy targets arising from alternative splicing
Abstract
Alternative pre-mRNA splicing (AS) is a prevalent mechanism and a main source of transcriptomic and proteomic complexity in cells. Dysregulation of AS is widespread in tumor transcriptomes. Cancer immunotherapies have transformed the treatment of aggressive tumors, but identification of novel tumor antigens remains challenging. Petabytes of sequencing data in public domains presents unprecedented opportunities to exploit AS-derived peptides as a new category in the tumor antigen repertoire. In this dissertation, novel computational methods were developed to detect AS variations in cancers with significant biological or therapeutic implications. Utilizing these new tools, we demonstrated that we can characterize the key AS changes responding to oncogenic signals alterations, and more importantly, systematically identify splicing events that are potential tumor antigens for targeted immunotherapies.
The first part of the dissertation describes Pathway Enrichment-Guided Activity Study of Alternative Splicing (PEGASAS), a novel computational framework identifying key splicing changes associated with oncogenic signals during disease progression from large-scale RNA-seq data. Although aberrant AS are widely detected in cancer, causes and consequences of AS dysregulations during cancer progression remain elusive. PEGASAS uses a pathway-guided approach for examining the effects of oncogenic signaling on splicing. Applying it to study a comprehensive prostate cancer dataset, we identified a conserved set of AS events regulated by oncogenic pathways and establish a role for Myc in regulating RNA processing. PEGASAS provides a generic framework to connect AS changes with a wide range of oncogenic alterations in cancers.
The second part of the dissertation presents Isoform peptides from RNA splicing for Immunotherapy target Screening (IRIS), a big data computational platform that integrates massive transcriptomic data along with proteomics data to characterize AS- derived tumor antigens for cancer immunotherapy. Exiting frameworks of tumor antigen discovery are predominantly somatic mutation-based, leaving AS-derived targets largely unexploited. IRIS employs a comprehensive reference panel that determines tumor AS events by leveraging splicing patterns from tens of thousands normal and tumor transcriptomes. Applying IRIS to analyze RNA-Seq data from 22 glioblastomas from patients, we identified candidate epitopes and validated their recognition by patient T cells. This work demonstrates IRIS’s utility for expanding targeted cancer immunotherapy by enabling big data-informed discoveries of a variety of AS-derived tumor antigens.
Main Content
Enter the password to open this PDF file:
-
-
-
-
-
-
-
-
-
-
-
-
-
-