The development of immunotherapies has revolutionized the treatment landscape for several cancers over the last decade, in particular for advanced melanoma:
exceptional clinical outcomes with durable survival are now obtained with anti-PD-1 or anti-PD-L1 therapies for a substantial number of patients with advanced melanoma while the median overall survival used to be less than 1 year. Yet, less than half of patients respond to immunotherapy and treatments can result in severe and long-lasting toxicities, and/or acquired resistance. One major challenge in cancer treatment is the inherent cellular heterogeneity that is part and parcel of tumorigenesis as much as drug resistance. Thus an accurate assessment of tumor heterogeneity, whilst still a major challenge, is essential for the development of effective therapies, specific and reliable biomarkers to predict drug response as well as to indicate which patient would benefit from double yet heavy immunotherapy.
Single-cell RNA-sequencing (scRNA-seq) emerged as a breakthrough technology to characterize the human tissue cellular heterogeneity including cancer biopsies. Compared to bulk RNA-Seq, this technology however suffers from strong technical variation and bias due to low coverage, extensive PCR amplification and low capture efficiency leading to high dropout rate. Furthermore most scRNA-seq protocols are typically heavily biased toward the 3’ end. As a result, most single cell studies have restricted analysis of transcriptome variation to the gene level. Yet alternative isoforms generated by alternative splicing (AS) and alternative polyadenylation (APA) are strong determinants of cell identity, cancer development, and drug/therapy resistance. Thus simultaneous assessment of gene expression and AS/APA from scRNA-seq holds great promise to efficiently tease apart the cellular heterogeneity in seemingly homogenous tumor cell populations. Because splicing analysis from scRNA-seq is still in its infancy, existing tools have not yet been benchmarked in an independent study or characterised for their ability to uncover rare cell populations, and best practices for performing exon-centric splicing analysis using single-cell data have to be established.
Here we propose to develop a framework to enable easy implementation and evaluation of AS/APA methods to scRNA-seq and then test their utility to identify rare tumor subpopulation in advanced melanoma. Our study will enable high-resolution reconstruction of cell population by combining gene expression analysis with aberrant mRNA metabolism enabling patient stratification and biomarker discovery. It will also shed light into the potential role of aberrant mRNA metabolism in cancer with potential for biomarkers discovery. Altogether this study will enable the development of methods that will enable to uncover the molecular mechanisms underlying melanoma pathogenesis, improve diagnostics and accelerate the identification of prognostic as well as treatment-predictive biomarkers.