From bulk RNA sequencing to spatial transcriptomics: a comparative review of differential gene expression analysis methods

BACKGROUND: Transcriptome analysis is essential to dissect the molecular mechanisms underlying phenotypic differences and disease mechanisms. Traditionally, microarrays were used, but RNA sequencing (RNA-seq) has become a more powerful and reproducible alternative. RNA-seq is widely applied for differential expression analysis (DEA) analysis, making it possible to characterize gene expression heterogeneity across diverse biological conditions. MAIN BODY: Several approaches are used in RNA-seq, including bulk RNA-seq, single-cell RNA sequencing (scRNA-seq), direct RNA sequencing (DRS), and spatial transcriptomics (ST), every distinct offering unique insight. Bulk RNA-seq offers a global perspective on transcriptional activity, while scRNA-seq dissects the intrinsic cellular diversity and developmental trajectories, aiding for delineating molecular profiles of low-abundance cell phenotypes and complex biology. However, both methods face challenges such as biases from short-read sequencing, limited isoform resolution, and the loss of spatial information. Advanced methods such as DRS and ST have emerged to address these limitations. RNA-seq data analysis comprises a multi-phase workflow, including read trimming, alignment, quantification, normalization, and differential expression testing. As sequencing technologies evolve, numerous computational tools and statistical methods have been developed to support DEA across bulk, scRNA-seq, DRS, and ST datasets, enhancing our ability to interpret transcriptomic changes in health and disease. Applications of RNA-seq include interrogation of multifactorial disease mechanisms, biomarker discovery, and the investigation of gene regulatory networks. CONCLUSION: Examining gene expression through these diverse RNA-seq modalities deepens our understanding of cellular processes, transcript modifications, tissue spatial organization, and gene expression variability. This review highlights current DEA methods in R and Python for bulk RNA-seq, scRNA-seq, DRS, and ST, showcasing the rapidly evolving landscape of transcriptome research.