pbmc scrna-seq dataset  (10X Genomics)

Journal: bioRxiv

Article Title: CSFeatures improves identification of cell type-specific differential features in single-cell and spatial omics data

doi: 10.1101/2025.05.21.655244

Figure Lengend Snippet: ( A ) Application scenarios of differential expression analysis in scRNA-seq data analysis. ( B ) The violin plot shows the expression distribution of the top-ranked differentially expressed gene (DEG), DCN , identified by the Wilcoxon rank-sum test for the fibroblast population in the gastric cancer dataset. While highly expressed in fibroblasts, DCN also exhibits significant expression in mural cells. ( C ) The scatter plots visualize the expression distribution of the gene DCN across different cell populations within the gastric cancer dataset. ( D ) The violin plot illustrates the expression distribution of the gene CD74 , identified by the Wilcoxon rank-sum test, for the B cell population in the PBMC dataset. CD74 is highly expressed in B cells as well as in several other cell populations. ( E ) The scatter plots demonstrate the distribution of CD74 and FAM177B gene expression across different cell populations in the PBMC dataset. ( F ) The violin plot illustrates the expression distribution of the top-ranked DEG, FAM177B , for the B cell population in the PBMC dataset. Identified through fold change ranking after filtering with P < 0.01, FAM177B exhibits low expression across all cell populations, including the B cell population.

Article Snippet: The PBMC scRNA-seq dataset can be downloaded at https://cf.10xgenomics.com/ samples/cell/pbmc3k.

Techniques: Quantitative Proteomics, Expressing, Gene Expression

Journal: bioRxiv

Article Title: CSFeatures improves identification of cell type-specific differential features in single-cell and spatial omics data

doi: 10.1101/2025.05.21.655244

Figure Lengend Snippet: ( A ) Overview of CSFeatures. CSFeatures takes a gene expression matrix and cell population labels as input, and computes a cell-to-cell correlation matrix following dimensionality reduction. For each gene, CSFeatures fully considers its expression level, the smoothness of its expression distribution, and the proportion of cells expressing the gene across all cell populations. Genes are ranked by their EI values, prioritizing those with strong cell type specificity. ( B ) For the CD8 T cell population in the PBMC dataset, the expression distribution of the top three genes identified by the Wilcoxon rank-sum test (top) and CSFeatures (bottom). ( C ) The bubble plots display the top five differentially expressed genes for each cell population identified by the Wilcoxon rank-sum test (left) and CSFeatures (right) in the human glioblastoma data. Colors represent expression levels, and bubble sizes correspond to the proportion of cells expressing each gene.

Article Snippet: The PBMC scRNA-seq dataset can be downloaded at https://cf.10xgenomics.com/ samples/cell/pbmc3k.

Techniques: Gene Expression, Expressing

Journal: bioRxiv

Article Title: CSFeatures improves identification of cell type-specific differential features in single-cell and spatial omics data

doi: 10.1101/2025.05.21.655244

Figure Lengend Snippet: ( A ) The top 10 differentially expressed genes (DEGs) identified by the Wilcoxon rank-sum test (left) and CSFeatures (right) for the B cell population in the PBMC dataset. ( B ) GO enrichment of the top 20 unique DEGs identified by the Wilcoxon rank-sum test (left) and CSFeatures (right) for the B cell population in the PBMC dataset. ( C ) For the endothelial cell population in the lung cancer dataset, the top downregulated genes in tumor tissue, TIMP1 and CRHBP , respectively identified by the Wilcoxon rank-sum test (left) and CSFeatures (right), are shown. ( D ) The scatter plots further show the expression distribution of TIMP1 and CRHBP between the groups, with color intensity representing expression levels.

Article Snippet: The PBMC scRNA-seq dataset can be downloaded at https://cf.10xgenomics.com/ samples/cell/pbmc3k.

Techniques: Expressing

Journal: International Journal of Molecular Sciences

Article Title: Identification of Human Global, Tissue and Within-Tissue Cell-Specific Stably Expressed Genes at Single-Cell Resolution

doi: 10.3390/ijms231810214

Figure Lengend Snippet: Comparison of SEGs identified from different datasets. ( A ) Comparison of the spatial-temporal SEGs (Microwell-Seq) and PBMC SEGs identified from 10x Genomics scRNA-Seq datasets. ( B ) Comparison of the Microwell-Seq SEG, Smart-Seq SEG and bulky RNA-seq HKG lists . ( C ) GESI heatmap of representative genes inconsistent among the Microwell-Seq SEG, Smart-Seq SEG and bulky RNA-seq HKG lists. Ten Microwell-Seq SEGs, five Smart-Seq SEGs and five bulky RNA-seq HKGs not covered by the other two gene lists (from the specific subsets shown in B) were selected randomly, and their cellcluster GESIs in two representative samples were shown. The samples were randomly selected from , both of which were from adult tissues. ( D ) Expression of two representative SEGs/HKGs specifically detected by Microwell-Seq ( RPL7 and RPS29 ), Smart-Seq ( CHD1 and SRPK1 ) or bulky RNA-seq ( TUBD1 and ERCC5 ) in single-cell clusters of a representative sample. The representative sample (an adult lung sample) was selected from randomly.

Article Snippet: To balance the power and precision, we tested the different combination of the parameters using a 10K PBMC scRNA-seq dataset from the 10x Genomics website, i.e., top 100, 500 and 1000 stably expressed genes within clusters, common within-cluster SEGs among 50%, 75% and 100% clusters, and an inter-cluster GESI cutoff of 0.667, 0.714, 0.769, 0.833 and 0.909.

Techniques: Comparison, RNA Sequencing, Expressing

Journal: International Journal of Molecular Sciences

Article Title: Identification of Human Global, Tissue and Within-Tissue Cell-Specific Stably Expressed Genes at Single-Cell Resolution

doi: 10.3390/ijms231810214

Figure Lengend Snippet: SEGs from tissues under stresses or with diseases. ( A ) Comparison of SEGs from PBMCs (or T cells for HIV-1 infection and control) of healthy donors and patients with diseases. ( B ) Comparison of SEGs from different healthy and diseased solid tissues. For both ( A , B ), the percentages of common SEGs (intersects of the Venn graphs) in healthy tissues were shown in the bar plots.

Article Snippet: To balance the power and precision, we tested the different combination of the parameters using a 10K PBMC scRNA-seq dataset from the 10x Genomics website, i.e., top 100, 500 and 1000 stably expressed genes within clusters, common within-cluster SEGs among 50%, 75% and 100% clusters, and an inter-cluster GESI cutoff of 0.667, 0.714, 0.769, 0.833 and 0.909.

Techniques: Comparison, Infection, Control

Journal: International Journal of Molecular Sciences

Article Title: Identification of Human Global, Tissue and Within-Tissue Cell-Specific Stably Expressed Genes at Single-Cell Resolution

doi: 10.3390/ijms231810214

Figure Lengend Snippet: Performance of the SEGdecon model. SEGdcon was used to predict the proportion of major immune cells in PBMC samples. Each testing dataset was shown in a column. The real and predicted composition percentages of each major immune cell type were shown in the pie charts. The average accuracy was shown at the bottom for each testing dataset.

Article Snippet: To balance the power and precision, we tested the different combination of the parameters using a 10K PBMC scRNA-seq dataset from the 10x Genomics website, i.e., top 100, 500 and 1000 stably expressed genes within clusters, common within-cluster SEGs among 50%, 75% and 100% clusters, and an inter-cluster GESI cutoff of 0.667, 0.714, 0.769, 0.833 and 0.909.

Techniques:

Journal: International Journal of Molecular Sciences

Article Title: Identification of Human Global, Tissue and Within-Tissue Cell-Specific Stably Expressed Genes at Single-Cell Resolution

doi: 10.3390/ijms231810214

Figure Lengend Snippet: Identification of SEGs at single-cell level. ( A ) The pipeline for SEG identification. ( B ) 408 spatial-temporal SEGs were identified from both adult and fetal tissues. ( C ) Expression of a representative spatial-temporal SEG, EIF1 , in single-cell clusters of representative samples. One sample was selected randomly for each of the adult lung, adult stomach, fetal lung and fetal stomach tissues, respectively . The single cells were clustered according to the gene expression profile, and shown as points in a two-dimension plane of Uniform Manifold Approximation and Projection (UMAP) . The gene expression level was represented by the color shown in the scale bar. ( D ) GESI heatmap of representative SEGs and non-SEGs in cell clusters of representative samples. Ten spatial-temporal SEGs and ten non-SEGs were randomly selected. Two scRNA-seq samples were selected from the adult tissues randomly . ( E ) Expression correlation of SEGs and non-SEGs between samples. All 408 SEGs and paired 408 non-SEGs randomly selected from the non-SEG gene set were used for the analysis and comparison. The Pearson Correlation Coefficients (PCCs) were calculated for SEGs and non-SEGs between each inter-sample cell-cluster pair, respectively, and there were 579,246 pairs in total. A Mann-Whitney U test was performed between the PCCs of SEGs and non-SEGs. The correlation of SEGs and non-SEGs between a representative cell-cluster pair was also shown.

Article Snippet: Other 10x Genomics PBMC scRNA-seq datasets including one from healthy subject (HC1), one from a TB patient (TB1) and another from a LTBI patient (LTBI1) were downloaded from NCBI GEO database, and the M, B and NK/T cells were annotated according to the original reference [ ].

Techniques: Expressing, MANN-WHITNEY

Journal: International Journal of Molecular Sciences

Article Title: Identification of Human Global, Tissue and Within-Tissue Cell-Specific Stably Expressed Genes at Single-Cell Resolution

doi: 10.3390/ijms231810214

Figure Lengend Snippet: Comparison of SEGs identified from different datasets. ( A ) Comparison of the spatial-temporal SEGs (Microwell-Seq) and PBMC SEGs identified from 10x Genomics scRNA-Seq datasets. ( B ) Comparison of the Microwell-Seq SEG, Smart-Seq SEG and bulky RNA-seq HKG lists . ( C ) GESI heatmap of representative genes inconsistent among the Microwell-Seq SEG, Smart-Seq SEG and bulky RNA-seq HKG lists. Ten Microwell-Seq SEGs, five Smart-Seq SEGs and five bulky RNA-seq HKGs not covered by the other two gene lists (from the specific subsets shown in B) were selected randomly, and their cellcluster GESIs in two representative samples were shown. The samples were randomly selected from , both of which were from adult tissues. ( D ) Expression of two representative SEGs/HKGs specifically detected by Microwell-Seq ( RPL7 and RPS29 ), Smart-Seq ( CHD1 and SRPK1 ) or bulky RNA-seq ( TUBD1 and ERCC5 ) in single-cell clusters of a representative sample. The representative sample (an adult lung sample) was selected from randomly.

Article Snippet: Other 10x Genomics PBMC scRNA-seq datasets including one from healthy subject (HC1), one from a TB patient (TB1) and another from a LTBI patient (LTBI1) were downloaded from NCBI GEO database, and the M, B and NK/T cells were annotated according to the original reference [ ].

Techniques: RNA Sequencing Assay, Expressing