Comparison of scRNA-seq data analysis method combinations
Briefings in Functional Genomics · 17.11.2022 · 21 (6) · 433-440
DOI: 10.1093/bfgp/elac027 ↗Kurz gefasst
Fachartikel in der Zeitschrift Briefings in Functional Genomics, online erschienen am 16.09.2022. Verfasst von Li Xu und 3 weiteren Autor:innen. Hauptthemen laut MEDLINE: Single-Cell Analysis, Data Analysis. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 37 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Fundamental Research, Fundamental Research Funds for the Central Universities.
Abstract (Originalsprache)
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.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Substanzen
Autor:innen und Institutionen
ORCID-Profile: Li Xu
Förderung
- National Natural Science Foundation of China
- Fundamental Research
- Fundamental Research Funds for the Central Universities
Zitieren
APA
Xu, L., Xue, T., Ding, W., & Shen, L. (2022). Comparison of scRNA-seq data analysis method combinations. Briefings in Functional Genomics, 21(6), 433-440. https://doi.org/10.1093/bfgp/elac027
Vancouver
Xu L, Xue T, Ding W, Shen L. Comparison of scRNA-seq data analysis method combinations. Briefings in Functional Genomics. 2022;21(6):433-440. doi:10.1093/bfgp/elac027
BibTeX
@article{Xu2022,
author = {Xu, Li and Xue, Tong and Ding, Weiyue and Shen, Linshan},
title = {Comparison of scRNA-seq data analysis method combinations},
journal = {Briefings in Functional Genomics},
year = {2022},
volume = {21},
number = {6},
pages = {433--440},
doi = {10.1093/bfgp/elac027}
}
EUVIMED