Abstract:Single-cell ribonucleic acid (RNA)-sequencing (scRNA-seq) data analysis refers to the use of appropriate methods to analyze the dataset generated by RNA-sequencing performed on the single-cell transcriptome. It usually contains three steps: normalization to eliminate the technical noise, dimensionality reduction to facilitate visual understanding and data compression and clustering to divide the data into several similarity-based clusters. In addition, the gene expression data contain a large number of zero counts. These zero counts are considered relevant to random dropout events induced by multiple factors in the sequencing experiments, such as low RNA input, and the stochastic nature of the gene expression pattern at the single-cell level. The zero counts can be eliminated only through the analysis of the scRNA-seq data, and although many methods have been proposed to this end, there is still a lack of research on the combined effect of existing methods. In this paper, we summarize the two kinds of normalization, two kinds of dimension reduction and three kinds of clustering methods widely used in the current mainstream scRNA-seq data analysis. Furthermore, we propose to combine these methods into 12 technology combinations, each with a whole set of scRNA-seq data analysis processes. We evaluated the proposed combinations using Goolam, a publicly available scRNA-seq, by comparing the final clustering results and found the most suitable collection scheme of these classic methods. Our results showed that using appropriate technology combinations can improve the efficiency and accuracy of the scRNA-seq data analysis. The combinations not only satisfy the basic requirements of noise reduction, dimension reduction and cell clustering but also ensure preserving the heterogeneity of cells in downstream analysis. The dataset, Goolam, used in the study can be obtained from the ArrayExpress database under the accession number E-MTAB-3321.

Accuracy, robustness and scalability of dimensionality reduction methods for single-cell RNA-seq analysis

A Comparison for Dimensionality Reduction Methods of Single-Cell RNA-seq Data

Dimension Reduction and Clustering Models for Single-Cell RNA Sequencing Data: A Comparative Study

Comparative Research of Different Dimension Reduction Methods Combined with RWR Network Smoothing in Single Cell RNA-seq Data

A quantitative framework for evaluating single-cell data structure preservation by dimensionality reduction techniques

A comprehensive survey of dimensionality reduction and clustering methods for single-cell and spatial transcriptomics data

Dimensionality Reduction for Single Cell RNA Sequencing Data Using Constrained Robust Non-Negative Matrix Factorization

Joint dimension reduction and clustering analysis of single-cell RNA-seq and spatial transcriptomics data

Joint learning dimension reduction and clustering of single-cell RNA-sequencing data

Comparative analysis of dimension reduction methods for cytometry by time-of-flight data

Comprehensive Evaluation of Noise Reduction Methods for Single-Cell RNA Sequencing Data

Comparison of Scrna-Seq Data Analysis Method Combinations.

Dimensionality Reduction of Single-Cell RNA Sequencing Data by Combining Entropy and Denoising AutoEncoder.

SSNMDI: a novel joint learning model of semi-supervised non-negative matrix factorization and data imputation for clustering of single-cell RNA-seq data

Optimal distance metrics for single-cell RNA-seq populations

A Data-Driven Clustering Recommendation Method for Single-Cell RNA-Sequencing Data

Feature selection and dimension reduction for single-cell RNA-Seq based on a multinomial model

A Streamlined Scrna-Seq Data Analysis Framework Based on Improved Sparse Subspace Clustering

ScDA: A Denoising AutoEncoder Based Dimensionality Reduction for Single-cell RNA-seq Data

Cluster Analysis of High-Dimensional scRNA Sequencing Data

Benchmark and Parameter Sensitivity Analysis of Single-Cell RNA Sequencing Clustering Methods