Single-cell analysis of alternative splicing in normal and malignant stem/progenitor cells

Thesis

Alternative splicing is the process of utilising a different combination of exons to generate different isoforms from the same gene. Therefore, alternative splicing represents an additional and underappreciated layer of complexity underlying gene expression. To date, high-throughput single-cell RNA-sequencing (scRNA-seq) analyses have focused on characterising gene expression programme, whereas alternative splicing remains challenging to investigate. One possible reason for the sparsity of single-cell alternative splicing studies is the lack of single-cell alternative splicing analytical frameworks. To this end, we have developed analytical frameworks to comprehensively capture the alternative splicing landscape in health and disease models, and to prioritise actionable spliced genes for downstream experimental studies.

The developed frameworks consist of MARVEL, VALERIE, and IMPACT. MARVEL is an R package that provides comprehensive functionalities for the detection and quantification of alternative splicing events to enable dimension reduction analysis, differential splicing analysis, and functional annotation of differentially spliced genes. Functional annotation features include biological pathway enrichment analysis and nonsense-mediated decay prediction. VALERIE is an R package for visual-based validation of differentially spliced genes identified from MARVEL. IMPACT is an integrated in-house database consisting of a collection of pre-processed publicly available myeloid neoplasm and cancer cell line data for prioritising clinically relevant and druggable spliced genes validated by VALERIE.

We validated and demonstrated the application of our analytical frameworks on scRNA-seq data generated from both plate- (e.g., Smart-seq2) and droplet- (e.g., 10x Genomics) based library preparation methods derived from homogeneous cell lines and heterogeneous haematopoietic stem and progenitor cells in health and disease states. We believe our analytical frameworks will be advantageous to biologists to reveal novel biological insights from scRNA-seq data.

Authors: Wen, S

• DOI: 10.5287/ora-gwk89wog4
• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English
• Keywords: single-cell; splicing; next-generation sequencing; cancer; genetics
• Subjects: High-throughput nucleotide sequencing; Cancer; Genetics