Accuracy, Robustness and Scalability of Dimensionality Reduction Methods for Single Cell RNAseq Analysis

Abstract

Background: Dimensionality reduction (DR) is an indispensable analytic component for many areas of single cell RNA sequencing (scRNAseq) data analysis. Proper DR can allow for effective noise removal and facilitate many downstream analyses that include cell clustering and lineage reconstruction. Unfortunately, despite the critical importance of DR in scRNAseq analysis and the vast number of DR methods developed for scRNAseq studies, however, few comprehensive comparison studies have been performed to evaluate the effectiveness of different DR methods in scRNAseq.Results: Here, we aim to fill this critical knowledge gap by providing a comparative evaluation of a variety of commonly used DR methods for scRNAseq studies. Specifically, we compared 18 different DR methods on 30 publicly available scRNAseq data sets that cover a range of sequencing techniques and sample sizes. We evaluated the performance of different DR methods for neighborhood preserving in terms of their ability to recover features of the original expression matrix, and for cell clustering and lineage reconstruction in terms of their accuracy and robustness. We also evaluated the computational scalability of different DR methods by recording their computational cost. Conclusions:Based on the comprehensive evaluation results, we provide important guidelines for choosing DR methods for scRNAseq data analysis. We also provide all analysis scripts used in the present study at www.xzlab.org/reproduce.html. Together, we hope that our results will serve as an important practical reference for practitioners to choose DR methods in the field of scRNAseq analysis. our results can serve as an important guideline for practitioners to choose DR methods in the field of scRNAseq analysis.

Tuning parameters of dimensionality reduction methods for single-cell RNA-seq analysis

“…type identification as a downstream task of DR. Both measures are frequently used to assess the performance of cell embedding techniques [18,11,20,23,10]. Another standard measure to assess the ability of a clustering algorithm to recover known classes is the adjusted rand index (ARI) [11,20,23,10]; however, we found that AMI and ARI are extremely correlated (Additional file 1: Fig.…”

“…Many computational methods have therefore been developed in recent years to take into account the specificities of scRNA-seq data and address the issues of data normalization, cell type identification, differential gene expression analysis, cell hierarchy reconstruction, or gene regulatory network inference (see [8,9] for recent reviews). In order to help practitioners choose an analysis pipeline among the many available, several studies have benchmarked algorithms and softwares for applications such as dimensionality reduction [10], clustering [11,12], differential expression [13], or trajectory inference [14].…”

Eleven grand challenges in single-cell data science

“…Some methods, especially those using deep learning, depend strongly on choice of hyperparameters [94]. There, more detailed comparisons that explore parameter spaces would be helpful, extending work like that from Sun et al [95] comparing dimensionality reduction methods. Such detailed benchmarking would also help to establish when normalization methods derived from explicit count models (e.g., [96,97]) may be preferable to imputation.…”

A survey of dimension reduction and classification methods for RNA-Seq data on malaria vector

“…Dimensionality reduction is very beneficial and vital when evaluating high dimensional data [57], it is an essential analytic factor for RNA-seq data analysis. Appropriate dimensionality reduction algorithms can facilitate effective evaluations and classification performance of different approaches in relation to their capability to improve features of the innovative expression in terms of their performance metrics such as accuracy, sensitivity, specificity, recall, robustness, computational scalability, computational cost, among others [58].…”