Abstract :

ABSTRACT

Single-cell RNA sequencing (scRNA-seq) enables high-resolution analysis of cellular heterogeneity, but traditional cell isolation methods like flow cytometry and laser microdissection often suffer from limitations in efficiency, viability, and bias. To overcome these challenges, computational tissue deconvolution approaches have emerged as effective alternatives. In this work, we introduce a high-performance computational pipeline for scRNA-seq data analysis that identifies and segregates cell populations based on marker gene expression. Our method incorporates advanced preprocessing, normalization, and clustering techniques, optimized for scalability and reproducibility in high-performance computing (HPC) environments. Compared to related tools, our pipeline offers enhanced adaptability across diverse datasets and experimental settings. We validated its performance using zebrafish ventricular tissue post-injury, effectively identifying key regenerative cell types such as immune cells, including macrophages. This approach supports in-depth biological discovery without prior physical cell separation and expands the potential of scRNA-seq applications in regenerative biology, immunology, and single-cell transcriptomics.

KEYWORDS

Single-cell RNA sequencing (scRNA-seq); Cell characterization; Computational deconvolution; High-performance computing (HPC); Marker gene expression; Zebrafish heart regeneration; Immune cell profiling; Macrophages; Tissue heterogeneity; Unbiased cell type identification.