Abstract:Single-cell RNA sequencing (scRNA-seq) enables the study of cellular diversity at single cell level. It provides a global view of cell-type specification during the onset of biological mechanisms such as developmental processes and human organogenesis. Various statistical, machine and deep learning-based methods have been proposed for cell-type classification. Most of the methods utilizes unsupervised lower dimensional projections obtained from for a large reference data. In this work, we proposed a reference-based method for cell type classification, called EnProCell. The EnProCell, first, computes lower dimensional projections that capture both the high variance and class separability through an ensemble of principle component analysis and multiple discriminant analysis. In the second phase, EnProCell trains a deep neural network on the lower dimensional representation of data to classify cell types. The proposed method outperformed the existing state-of-the-art methods when tested on four different data sets produced from different single-cell sequencing technologies. The EnProCell showed higher accuracy (98.91) and F1 score (98.64) than other methods for predicting reference from reference datasets. Similarly, EnProCell also showed better performance than existing methods in predicting cell types for data with unknown cell types (query) from reference datasets (accuracy:99.52; F1 score: 99.07). In addition to improved performance, the proposed methodology is simple and does not require more computational resources and time. the EnProCell is available at <a class="link-external link-https" href="https://github.com/umar1196/EnProCell" rel="external noopener nofollow">this https URL</a>.

What problem does this paper attempt to address?

The paper attempts to address the issue of accuracy in cell type classification within single-cell RNA sequencing (scRNA-seq) data. Specifically, the authors propose a new reference data-driven method called EnProCell to improve cell type classification performance from single-cell RNA sequencing data. ### Background and Problem Single-cell RNA sequencing technology (scRNA-seq) enables researchers to study cell diversity at the single-cell level, providing a global view of cell type specificity in biological mechanisms such as developmental processes and human organogenesis. However, the data generated from scRNA-seq experiments are typically high-dimensional and sparse, posing challenges for cell type classification. Existing methods mostly rely on unsupervised low-dimensional projections, which, while capturing high variance in the data, do not necessarily ensure class separability, thereby affecting classification performance. ### Solution To overcome this issue, the authors propose the EnProCell method. The main steps of this method include: 1. **Low-Dimensional Projection**: Combining Principal Component Analysis (PCA) and Multiple Discriminant Analysis (MDA) to obtain low-dimensional projections. PCA is used to capture components with high variance, while MDA ensures class separability. 2. **Deep Neural Network Training**: Training a deep neural network on the low-dimensional representation of the data to classify cell types. ### Experimental Results The authors tested the EnProCell method on multiple datasets and compared it with existing state-of-the-art methods. The results show that EnProCell outperforms other methods on several metrics, particularly in terms of classification performance on both reference and query datasets. Specifically: - **Reference Dataset Classification**: EnProCell achieved higher accuracy and F1 scores than other methods on datasets such as PBMC1, Baron, and Muraro. - **Query Dataset Classification**: EnProCell also excelled in predicting unknown cell types in query datasets, significantly improving accuracy and F1 scores. ### Main Contributions 1. **Improved Low-Dimensional Projection**: By combining PCA and MDA, EnProCell better captures high variance and class separability in the data. 2. **Efficient Classification Performance**: EnProCell demonstrates excellent classification performance across multiple datasets without requiring excessive computational resources and time. 3. **Broad Applicability**: The method is not only suitable for reference datasets but also effectively handles query datasets, showing high generalization capability. Overall, the paper addresses the accuracy issue in cell type classification within single-cell RNA sequencing data by proposing the EnProCell method, providing new tools and insights for research in the related field.

Lower-dimensional projections of cellular expression improves cell type classification from single-cell RNA sequencing

EnClaSC: a novel ensemble approach for accurate and robust cell-type classification of single-cell transcriptomes

XgCPred: Cell type classification using XGBoost-CNN integration and exploiting gene expression imaging in single-cell RNAseq data

Evaluation of some aspects in supervised cell type identification for single-cell RNA-seq: classifier, feature selection, and reference construction

Reference-free Cell-type Annotation for Single-cell Transcriptomics using Deep Learning with a Weighted Graph Neural Network

Identification of cell-type-specific genes in multimodal single-cell data using deep neural network algorithm

A Hybrid Deep Clustering Approach for Robust Cell Type Profiling Using Single-cell RNA-seq Data

Cell Type Identification from Single-Cell Transcriptomic Data Via Gene Embedding.

scDeepInsight: a supervised cell-type identification method for scRNA-seq data with deep learning

Abstract 3520: A scalable single cell RNA-seq pipeline leveraging machine learning and high-quality references for cell-type prediction

Recent Developments and Applications of Single-Cell RNA Sequencing Technology in Cell Classification

Autoencoder-based Cluster Ensembles for Single-Cell RNA-seq Data Analysis

Automated identification of Cell Types in Single Cell RNA Sequencing

Cell Type Differentiation Using Network Clustering Algorithms

Ranking of cell clusters in a single-cell RNA-sequencing analysis framework using prior knowledge

scAnno: a deconvolution strategy-based automatic cell type annotation tool for single-cell RNA-sequencing data sets

scDetect: a rank-based ensemble learning algorithm for cell type identification of single-cell RNA sequencing in cancer

EnDecon: cell type deconvolution of spatially resolved transcriptomics data via ensemble learning

Highly Accurate Estimation of Cell Type Abundance in Bulk Tissues Based on Single‐Cell Reference and Domain Adaptive Matching

Integration for single-cell RNA sequencing data based on the shared cell type assignment

Identification of cell types from single cell data using stable clustering