Bioinformatics approaches for the study of isoforms in single-cell transcriptomics coupled with long-read sequencing
Year: 2025
Type: dissertation
Abstract: Recent advances in long-read single-cell transcriptomics (scRNA-seq) have transformed the study of cellular hetero- geneity. These technologies overcome key limitations of short-read scRNA-seq methods, particularly for transcript cover- age, isoform detection, and novel transcript identification. The emergence of long-read scRNA-seq protocols now enables the characterization of full-length transcripts and the analysis of alternative splicing at single-cell resolution. However, when I began this thesis, long-read scRNA-seq approaches still had lower accuracy and sequencing depth than short- read methods, making data analysis challenging. Computational methods have since been developed for cell barcode (CB) and unique molecular identifier (UMI) error correction and for isoform quantification, but systematic benchmarking was missing.During my PhD, I conducted the first comprehensive benchmark of ten scRNA-seq long-read analysis methods, evaluating their performance in barcode and UMI assignment, gene expression quantification, and isoform detection across diverse datasets, protocols, and sequencing depths. In the absence of reliable ground truth and a dedicated simulation tool, I developed a simulation framework for Nanopore scRNA-seq and built a reusable benchmarking workflow that can integrate new tools and datasets. This work provides practical guidance for users and developers, and highlights factors influencing method performance.Finally, in collaboration with Pr. Topilko’s team (IMRB), I applied long-read scRNA-seq to study alternative splicing and isoform switching during malignant transformation in neurofibromatosis type 1 (NF1) tumors in mouse models. Using state-of-the-art tools (Sicelore 2.1, IsoQuant) and new analysis workflows, I identified recurrent isoform switches linked to cell cycle states and glial-to-mesenchymal transition in NF1 tumors.
Source: HAL
Author: Hamraoui, Ali
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7128649220