Clustering ensemble in scRNA-seq data analysis: Methods, applications and challenges

Comprehensive review on single-cell RNA sequencing: A new frontier in Alzheimer's disease research

Alzheimer's disease (AD) is a multifaceted neurodegenerative condition marked by gradual cognitive deterioration and the loss of neurons. While conventional bulk RNA sequencing techniques have shed light on AD pathology, they frequently obscure the cellular diversity within brain tissues. The advent of single-cell RNA sequencing (scRNA-seq) has transformed our capability to analyze the cellular composition of AD, allowing for the detection of unique cell populations, rare cell types, and gene expression alterations at an individual cell level. This review examines the use of scRNA-seq in AD research, focusing on its contributions to understanding cellular diversity, disease progression, and potential therapeutic targets. We discuss key technological innovations, data analysis techniques, and challenges associated with scRNA-seq in studying AD. Furthermore, we highlight recent studies that have utilized scRNA-seq to identify novel biomarkers, uncover disease-associated pathways, and elucidate the role of non-neuronal cells, such as microglia and astrocytes, in AD pathogenesis. By providing a comprehensive overview of advancements in scRNA-seq for unraveling cellular heterogeneity in AD, this review highlights the transformative impact of scRNA-seq on our comprehension of disease mechanisms and the creation of targeted treatments.

Single-cell transcriptome sequencing analysis reveals intra-tumor heterogeneity in esophageal squamous cell carcinoma

Esophageal squamous cell carcinoma (ESCC) is a prevalent malignant tumor of the digestive system that poses a significant threat to human life and health. It is crucial to thoroughly investigate the mechanisms of esophageal carcinogenesis and identify potential key molecular events in its carcinogenesis. Single-cell transcriptome sequencing is an emerging technology that has gained prominence in recent years for studying molecular mechanisms, which may help to further explore the underlying mechanisms of the ESCC tumor microenvironment in depth. The single-cell dataset was obtained from GSE160269 in the Gene Expression Omnibus database, including 60 tumor samples and four paracancer samples. The single-cell data underwent dimensional reduction clustering analysis to identify clusters and annotate expression profiles. Subcluster analysis was conducted for each cellular taxon. Copy number variation analysis of tumor cell subpopulations was performed to primarily identify malignant cells within them. A proposed chronological analysis was performed to obtain the process of cell differentiation. In addition, cell communication, transcription factor analysis, and tumor pathway analysis were also performed. Relevant risk models and key genes were established by univariate COX regression and LASSO analysis. The key genes obtained from the screen were subjected to appropriate silencing and cellular assays, including CCK-8, 5-ethynyl-2'-deoxyuridine, colony formation, and western blot. Single-cell analysis revealed that normal samples contained a large number of fibroblasts, T cells, and B cells, with fewer other cell types, whereas tumor samples exhibited a relatively balanced distribution of cell types. Subclassification analysis of immune cells, fibroblasts, endothelial cells, and epithelial cells revealed their specific spatial characteristics. The prognostic risk model, we constructed successfully, achieved accurate prognostic stratification for ESCC patients. The screened key gene, UPF3A, was found to be significantly associated with the development of ESCC by cellular assays. This process might be linked to the phosphorylation of ERK and P38. Single-cell transcriptome analysis successfully revealed the distribution of cell types and major expressed factors in ESCC patients, which could facilitate future in-depth studies on the therapeutic mechanisms of ESCC.

Variable selection for nonlinear dimensionality reduction of biological datasets through bootstrapping of correlation networks

Identifying the most relevant variables or features in massive datasets for dimensionality reduction can lead to improved and more informative display, faster computation times, and more explainable models of complex systems. Despite significant advances and available algorithms, this task generally remains challenging, especially in unsupervised settings. In this work, we propose a method that constructs correlation networks using all intervening variables and then selects the most informative ones based on network bootstrapping. The method can be applied in both supervised and unsupervised scenarios. We demonstrate its functionality by applying Uniform Manifold Approximation and Projection for dimensionality reduction to several high-dimensional biological datasets, derived from 4D live imaging recordings of hundreds of morpho-kinetic variables, describing the dynamics of thousands of individual leukocytes at sites of prominent inflammation. We compare our method with other standard ones in the field, such as Principal Component Analysis and Elastic Net, showing that it outperforms them. The proposed method can be employed in a wide range of applications, encompassing data analysis and machine learning.