Reliable and accurate gene expression quantification with subpopulation structure-aware constraints for single-cell RNA sequencing
Tu, C.-C.; Hung, J.-H.
Show abstract
BackgroundSingle-cell RNA sequencing (scRNA-seq) analysis analyzes the type and state of individual cells by estimating the gene expression of each cell and enables researchers to study the biological phenomena that cannot be observed in bulk RNA sequencing. MotivationHowever, the current scRNA-seq quantification tools estimate the gene expression profile of each cell independently, ignoring the fact that there are multiple cell types in the scRNA-seq data and the expression level should be highly correlated with the cell type. Since scRNA-seq suffers from a low sequencing depth, the conventional strategy leads to a high proportion of missing values in the gene expression profile, obscuring the biological characteristics of cell subpopulations and further impacting the correctness of the subsequent downstream analysis. ResultsIn this study, we proposed Quasic, a novel scRNA-seq quantification pipeline which examines the potential cell subpopulation information during quantification, and uses the information to calculate the gene expression level. Using the human peripheral blood mononuclear cells and the simulated doublet dataset, we verified that Quasic not only correctly reinforced the cell signatures, but also identified the corresponding cell subpopulations and biological pathways more accurately. In addition, we also applied Quasic to the breast cancer cell line dataset (MCF-7), and successfully identified more potentially therapeutic resistant cells of which characteristics are consistent with that from previous studies. ConclusionsThe proposed pipeline can let the gene expression profile of each cell be more consistent with the corresponding subpopulation, making the biological features unique to the subpopulation more apparent and convenient for analysis. By using Quasic, researchers can effectively extract the desired cell subpopulation information from their sampled cells, enable them to perform cell subpopulation-related studies more accurately.
Matching journals
The top 5 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Enhancing single-cell cellular state inference by incorporating molecular network features 98%
- scDeepInsight: a supervised cell-type identification method for scRNA-seq data with deep learning 97%
- SSMD: A semi-supervised approach for a robust cell type identification and deconvolution of mouse transcriptomics data 96%
Similar papers in this journal
Similar papers in this journal
- DeepCORE: An interpretable multi-view deep neural network model to detect co-operative regulatory elements 94%
- Finding new cancer epigenetic and genetic biomarkers from cell-free DNA by combining SALP-seq and machine learning：esophageal cancer as an example 94%
- Gra-CRC-miRTar: The pre-trained nucleotide-to-graph neural networks to identify potential miRNA targets in colorectal cancer 94%
Similar papers in this journal
- BERMUDA: A novel deep transfer learning method for single-cell RNA sequencing batch correction reveals hidden high-resolution cellular subtypes 96%
- scCDC: a computational method for gene-specific contamination detection and correction in single-cell and single-nucleus RNA-seq data 94%
- Accuracy, Robustness and Scalability of Dimensionality Reduction Methods for Single Cell RNAseq Analysis 94%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.