bio-single-cell-markers-annotation

name: bio-single-cell-markers-annotation description: Find marker genes and annotate cell types in single-cell RNA-seq using Seurat (R) and Scanpy (Python). Use for differential expression between clusters, identifying cluster-specific markers, scoring gene sets, and assigning cell type labels. Use when finding marker genes and annotating clusters. tool_type: mixed primary_tool: Seurat measurable_outcome: Execute skill workflow successfully with valid output within 15 minutes. allowed-tools: - read_file - run_shell_command

Marker Genes and Cell Type Annotation

Find differentially expressed genes between clusters and annotate cell types.

Scanpy (Python)

Required Imports

import scanpy as sc
import pandas as pd

Find Markers for All Clusters

# Find marker genes for each cluster vs all others
sc.tl.rank_genes_groups(adata, groupby='leiden', method='wilcoxon')

# View top markers
sc.pl.rank_genes_groups(adata, n_genes=10, sharey=False)

# Get results as DataFrame
markers = sc.get.rank_genes_groups_df(adata, group=None)
print(markers.head(20))

Marker Detection Methods

# Wilcoxon rank-sum test (default, recommended)
sc.tl.rank_genes_groups(adata, groupby='leiden', method='wilcoxon')

# t-test
sc.tl.rank_genes_groups(adata, groupby='leiden', method='t-test')

# Logistic regression
sc.tl.rank_genes_groups(adata, groupby='leiden', method='logreg')

Filter Markers

# Get markers with filters
markers = sc.get.rank_genes_groups_df(adata, group='0')
significant = markers[(markers['pvals_adj'] < 0.05) & (markers['logfoldchanges'] > 1)]
print(f'Cluster 0 significant markers: {len(significant)}')

# Filter all groups
sc.tl.filter_rank_genes_groups(adata, min_fold_change=1.5, min_in_group_fraction=0.25)

Compare Specific Clusters

# Find markers between two specific clusters
sc.tl.rank_genes_groups(adata, groupby='leiden', groups=['0'], reference='1', method='wilcoxon')
sc.pl.rank_genes_groups(adata, n_genes=10)

Visualize Marker Expression

# Dot plot of top markers per cluster
markers_to_plot = ['CD3D', 'CD8A', 'MS4A1', 'CD14', 'FCGR3A', 'NKG7']
sc.pl.dotplot(adata, var_names=markers_to_plot, groupby='leiden')

# Stacked violin
sc.pl.stacked_violin(adata, var_names=markers_to_plot, groupby='leiden')

# Heatmap
sc.pl.rank_genes_groups_heatmap(adata, n_genes=5, groupby='leiden')

# Matrix plot
sc.pl.matrixplot(adata, var_names=markers_to_plot, groupby='leiden')

Gene Set Scoring

# Score cells for gene set expression
t_cell_genes = ['CD3D', 'CD3E', 'CD4', 'CD8A', 'CD8B']
sc.tl.score_genes(adata, gene_list=t_cell_genes, score_name='T_cell_score')

# Visualize score
sc.pl.umap(adata, color='T_cell_score')

Cell Cycle Scoring

# Score cell cycle phases
s_genes = ['MCM5', 'PCNA', 'TYMS', 'FEN1', 'MCM2']  # S phase genes
g2m_genes = ['HMGB2', 'CDK1', 'NUSAP1', 'UBE2C', 'BIRC5']  # G2/M genes

sc.tl.score_genes_cell_cycle(adata, s_genes=s_genes, g2m_genes=g2m_genes)
sc.pl.umap(adata, color=['S_score', 'G2M_score', 'phase'])

Manual Cell Type Annotation

# Create annotation dictionary
cluster_annotations = {
    '0': 'CD4 T cells',
    '1': 'CD14 Monocytes',
    '2': 'B cells',
    '3': 'CD8 T cells',
    '4': 'NK cells',
    '5': 'FCGR3A Monocytes'
}

# Add annotations
adata.obs['cell_type'] = adata.obs['leiden'].map(cluster_annotations)

# Visualize
sc.pl.umap(adata, color='cell_type')

Export Markers

# Export all markers to CSV
markers = sc.get.rank_genes_groups_df(adata, group=None)
markers.to_csv('all_markers.csv', index=False)

# Export top markers per cluster
top_markers = markers.groupby('group').head(20)
top_markers.to_csv('top_markers.csv', index=False)

Seurat (R)

Required Libraries

library(Seurat)
library(dplyr)

Find All Markers

# Find markers for all clusters
all_markers <- FindAllMarkers(seurat_obj, only.pos = TRUE, min.pct = 0.25, logfc.threshold = 0.25)

# View top markers per cluster
top_markers <- all_markers %>%
    group_by(cluster) %>%
    slice_max(n = 5, order_by = avg_log2FC)
print(top_markers)

Find Markers for Specific Cluster

# Markers for cluster 0 vs all others
cluster0_markers <- FindMarkers(seurat_obj, ident.1 = 0, min.pct = 0.25)
head(cluster0_markers)

Compare Two Clusters

# Find markers between two specific clusters
markers_0_vs_1 <- FindMarkers(seurat_obj, ident.1 = 0, ident.2 = 1, min.pct = 0.25)
head(markers_0_vs_1)

Marker Detection Methods

# Wilcoxon (default, fast)
markers <- FindMarkers(seurat_obj, ident.1 = 0, test.use = 'wilcox')

# MAST (recommended for DE)
markers <- FindMarkers(seurat_obj, ident.1 = 0, test.use = 'MAST')

# DESeq2
markers <- FindMarkers(seurat_obj, ident.1 = 0, test.use = 'DESeq2')

# Logistic regression
markers <- FindMarkers(seurat_obj, ident.1 = 0, test.use = 'LR')

Visualize Markers

# Feature plot on UMAP
FeaturePlot(seurat_obj, features = c('CD3D', 'MS4A1', 'CD14', 'NKG7'))

# Violin plot
VlnPlot(seurat_obj, features = c('CD3D', 'MS4A1', 'CD14'))

# Dot plot
markers_to_plot <- c('CD3D', 'CD8A', 'MS4A1', 'CD14', 'FCGR3A', 'NKG7')
DotPlot(seurat_obj, features = markers_to_plot) + RotatedAxis()

# Heatmap
top10 <- all_markers %>%
    group_by(cluster) %>%
    top_n(n = 10, wt = avg_log2FC)
DoHeatmap(seurat_obj, features = top10$gene)

Gene Module Scoring

# Score cells for gene set
t_cell_genes <- list(c('CD3D', 'CD3E', 'CD4', 'CD8A', 'CD8B'))
seurat_obj <- AddModuleScore(seurat_obj, features = t_cell_genes, name = 'T_cell_score')

# Visualize
FeaturePlot(seurat_obj, features = 'T_cell_score1')

Cell Cycle Scoring

# Built-in cell cycle genes
s.genes <- cc.genes$s.genes
g2m.genes <- cc.genes$g2m.genes

seurat_obj <- CellCycleScoring(seurat_obj, s.features = s.genes, g2m.features = g2m.genes)
DimPlot(seurat_obj, group.by = 'Phase')

Manual Cell Type Annotation

# Rename cluster identities
new_cluster_ids <- c(
    '0' = 'CD4 T cells',
    '1' = 'CD14 Monocytes',
    '2' = 'B cells',
    '3' = 'CD8 T cells',
    '4' = 'NK cells',
    '5' = 'FCGR3A Monocytes'
)

seurat_obj <- RenameIdents(seurat_obj, new_cluster_ids)
DimPlot(seurat_obj, reduction = 'umap', label = TRUE)

# Store in metadata
seurat_obj$cell_type <- Idents(seurat_obj)

Export Markers

# Export to CSV
write.csv(all_markers, file = 'all_markers.csv', row.names = FALSE)

# Export top markers
write.csv(top_markers, file = 'top_markers.csv', row.names = FALSE)

Common PBMC Markers

Method Comparison

Related Skills

• clustering - Must cluster before finding markers
• preprocessing - Data must be normalized
• data-io - Export annotated data