metabolicpathwayactivity

Purpose

Calculates pathway activity scores for metabolic pathways across different cell groups and subsets. This process quantifies the metabolic activity of each pathway per group, generating visualizations (heatmaps and violin plots) to compare metabolic states between clusters or conditions. Based on the methodology from Xiao et al. (2019) Nature Communications.

When to Use

• Third step in metabolic workflow: After MetabolicInput and MetabolicExprImputation (optional)
• To quantify pathway-level metabolism: When you need scores for each metabolic pathway per group
• Compare metabolic states: To identify differences in pathway activity between clusters, treatments, or conditions
• Metabolic profiling visualization: When you need heatmaps showing pathway activity across groups and violin plots showing distribution
• Comprehensive metabolic analysis: As part of the ScrnaMetabolicLandscape group for complete metabolic landscape analysis

Configuration Structure

Process Enablement

MetabolicPathwayActivity is part of the ScrnaMetabolicLandscape group. Enable it by enabling the group:

[ScrnaMetabolicLandscape]
cache = true

Input Specification

MetabolicPathwayActivity receives input automatically from MetabolicInput:

[ScrnaMetabolicLandscape.in]
srtobj = ["SeuratClustering"]  # Input from upstream clustering process

Environment Variables

All configuration is done at the ScrnaMetabolicLandscape group level:

[ScrnaMetabolicLandscape.envs]
# Core configuration (inherited by all metabolic processes)
gmtfile = "KEGG_2021_Human"           # Metabolic pathways database
group_by = "seurat_clusters"            # Column to group cells (e.g., "cluster")
subset_by = "treatment"                 # Optional: Subset by metadata column
ncores = 1                              # Number of cores for parallelization

MetabolicPathwayActivity-Specific Configuration

[ScrnaMetabolicLandscape.MetabolicPathwayActivity.envs]
# Statistical analysis
ntimes = 5000                          # Number of permutations for p-value estimation

# Plot customization (default plots)
[ScrnaMetabolicLandscape.MetabolicPathwayActivity.envs.plots]
"Pathway Activity (violin plot)" = {
    plot_type = "violin",               # Options: "heatmap", "violin", "merged_heatmap"
    add_box = true,                     # Add box plot inside violin
    devpars = { res = 100 }             # Plot resolution
}
"Pathway Activity (heatmap)" = {
    plot_type = "heatmap",
    devpars = { res = 100 }
}
"All Subsets (merged)" = {
    plot_type = "merged_heatmap",       # All subsets in one plot
    devpars = { res = 100 }
}

# Multiple analysis cases (advanced)
[ScrnaMetabolicLandscape.MetabolicPathwayActivity.envs.cases]
"Treatment" = {
    subset_by = "treatment",            # Analyze by treatment groups
    group_by = "seurat_clusters",
    plots = { "Treatment Activity" = { plot_type = "violin", add_box = true } }
}
"Response" = {
    subset_by = "response",             # Analyze by response groups
    group_by = "seurat_clusters",
    plots = { "Response Activity" = { plot_type = "heatmap" } }
}

Pathway Activity Scoring

Scoring Method

MetabolicPathwayActivity uses AUCell-like scoring to calculate pathway activity:

• AUCell-like: Area Under the Curve calculation - ranks gene expression for each cell, then calculates area under the curve for genes in each pathway
• Permutation-based p-values: Uses ntimes permutations to estimate statistical significance
• Normalized scores: Scores are normalized to enable comparison across pathways and groups

Scoring Process

1. Gene ranking: For each cell, genes are ranked by expression level
2. Pathway AUC calculation: For each pathway, calculate AUC using the ranking and pathway gene list
3. Permutation testing: Randomly permute gene rankings ntimes times to estimate null distribution
4. P-value estimation: Compare observed AUC to null distribution to calculate significance
5. Score aggregation: Aggregate cell-level scores to group-level scores for visualization

GMT File Sources

The gmtfile parameter accepts:

• Built-in databases: "KEGG_2021_Human", "Reactome_Pathways_2024", "BioCarta_2016", "MSigDB_Hallmark_2020"
• Custom files: Local paths or URLs to GMT format files
• See /skills/processes/metabolicinput.md for detailed database options

Configuration Examples

Minimal Configuration (Default Settings)

[ScrnaMetabolicLandscape]
[ScrnaMetabolicLandscape.in]
srtobj = ["SeuratClustering"]

[ScrnaMetabolicLandscape.envs]
gmtfile = "KEGG_2021_Human"
group_by = "seurat_clusters"

Custom Plots with High Resolution

[ScrnaMetabolicLandscape]
[ScrnaMetabolicLandscape.envs]
gmtfile = "KEGG_2021_Human"
group_by = "seurat_clusters"

[ScrnaMetabolicLandscape.MetabolicPathwayActivity.envs]
ntimes = 10000  # More permutations for robust p-values

[ScrnaMetabolicLandscape.MetabolicPathwayActivity.envs.plots]
"Activity Violin" = {
    plot_type = "violin",
    add_box = true,
    devpars = { width = 1200, height = 800, res = 150 }
}
"Activity Heatmap" = {
    plot_type = "heatmap",
    devpars = { width = 1400, height = 1000, res = 150 }
}
"Merged Heatmap" = {
    plot_type = "merged_heatmap",
    devpars = { width = 1600, height = 1200, res = 150 }
}

Treatment Comparison Analysis

[ScrnaMetabolicLandscape]
[ScrnaMetabolicLandscape.envs]
gmtfile = "KEGG_2021_Human"
group_by = "seurat_clusters"
subset_by = "treatment"  # Compare between treatment groups

[ScrnaMetabolicLandscape.MetabolicPathwayActivity.envs]
ntimes = 5000
ncores = 4

[ScrnaMetabolicLandscape.MetabolicPathwayActivity.envs.plots]
"Treatment Violin" = { plot_type = "violin", add_box = true }
"Treatment Heatmap" = { plot_type = "heatmap" }

Multiple Analysis Cases (Advanced)

[ScrnaMetabolicLandscape]
[ScrnaMetabolicLandscape.envs]
gmtfile = "KEGG_2021_Human"
group_by = "seurat_clusters"
ncores = 8

# Case 1: Treatment analysis
[ScrnaMetabolicLandscape.MetabolicPathwayActivity.envs.cases.Treatment]
subset_by = "treatment"
group_by = "seurat_clusters"
ntimes = 5000
plots = {
    "Treatment Activity" = { plot_type = "violin", add_box = true, devpars = { res = 150 } }
}

# Case 2: Response analysis
[ScrnaMetabolicLandscape.MetabolicPathwayActivity.envs.cases.Response]
subset_by = "response"
group_by = "seurat_clusters"
ntimes = 10000
plots = {
    "Response Heatmap" = { plot_type = "heatmap", devpars = { res = 150 } }
}

Energy Metabolism Focus

[ScrnaMetabolicLandscape]
[ScrnaMetabolicLandscape.envs]
# Use custom GMT focused on energy pathways
gmtfile = "/data/pathways/energy_metabolism.gmt"
group_by = "seurat_clusters"
ncores = 4

[ScrnaMetabolicLandscape.MetabolicPathwayActivity.envs]
ntimes = 10000

[ScrnaMetabolicLandscape.MetabolicPathwayActivity.envs.plots]
"Energy Pathways" = { plot_type = "heatmap", devpars = { res = 150 } }

Common Patterns

Pattern 1: All Pathways with Default Settings

Standard analysis with KEGG pathways:

[ScrnaMetabolicLandscape]
[ScrnaMetabolicLandscape.envs]
gmtfile = "KEGG_2021_Human"
group_by = "seurat_clusters"

Pattern 2: Energy Metabolism Focus (Glycolysis + OXPHOS)

Focus on specific energy pathways using custom GMT:

[ScrnaMetabolicLandscape]
[ScrnaMetabolicLandscape.envs]
gmtfile = "/path/to/energy_pathways.gmt"
group_by = "seurat_clusters"

[ScrnaMetabolicLandscape.MetabolicPathwayActivity.envs]
ntimes = 5000

[ScrnaMetabolicLandscape.MetabolicPathwayActivity.envs.plots]
"Energy Activity" = { plot_type = "violin", add_box = true }

Pattern 3: Multiple Condition Comparison

Compare metabolic activity across different conditions:

[ScrnaMetabolicLandscape]
[ScrnaMetabolicLandscape.envs]
gmtfile = "Reactome_Pathways_2024"
group_by = "seurat_clusters"
subset_by = "condition"  # e.g., "control", "treatment_A", "treatment_B"

[ScrnaMetabolicLandscape.MetabolicPathwayActivity.envs]
ntimes = 10000

[ScrnaMetabolicLandscape.MetabolicPathwayActivity.envs.plots]
"Condition Comparison" = { plot_type = "heatmap" }
"Condition Distribution" = { plot_type = "violin", add_box = true }
"All Conditions" = { plot_type = "merged_heatmap" }

Pattern 4: High-Throughput Screening

For large datasets requiring parallel processing:

[ScrnaMetabolicLandscape]
[ScrnaMetabolicLandscape.envs]
gmtfile = "KEGG_2021_Human"
group_by = "seurat_clusters"
ncores = 16

[ScrnaMetabolicLandscape.MetabolicPathwayActivity.envs]
ntimes = 2000  # Fewer permutations for speed

Pattern 5: Publication-Quality Plots

High-resolution plots for manuscripts:

[ScrnaMetabolicLandscape]
[ScrnaMetabolicLandscape.envs]
gmtfile = "KEGG_2021_Human"
group_by = "seurat_clusters"

[ScrnaMetabolicLandscape.MetabolicPathwayActivity.envs]
ntimes = 10000

[ScrnaMetabolicLandscape.MetabolicPathwayActivity.envs.plots]
"Pathway Activity Heatmap" = {
    plot_type = "heatmap",
    devpars = { width = 1600, height = 1200, res = 300 }
}
"Pathway Activity Violin" = {
    plot_type = "violin",
    add_box = true,
    devpars = { width = 1200, height = 800, res = 300 }
}

Dependencies

Upstream Processes

• Required: MetabolicInput (part of ScrnaMetabolicLandscape group)
• Optional: MetabolicExprImputation (if imputation enabled with noimpute = false)
• Root: CombinedInput → requires SeuratClustering or similar clustering process

Downstream Processes

• Parallel: Runs alongside MetabolicFeatures and MetabolicPathwayHeterogeneity (same group)
• Optional: Can feed into visualization or reporting processes

Data Requirements

• Seurat object with normalized expression data
• Metadata column specified in group_by (e.g., cluster assignments)
• Optional metadata column in subset_by for subset analysis
• GMT file with metabolic pathway gene sets matching Seurat object gene names

Output Format

Output Files

MetabolicPathwayActivity generates the following outputs in the outdir directory (default: {{in.sobjfile | stem}}.pathwayactivity):

• Pathway activity scores: Tab-delimited files with pathway activity scores per group
• Heatmap plots: PNG/PDF images showing pathway activity heatmaps
• Violin plots: PNG/PDF images showing pathway activity distribution
• Merged heatmaps: Combined heatmaps for all subsets (if configured)

Score Interpretation

• Higher scores: Greater pathway activity in the group
• Lower scores: Lower pathway activity in the group
• P-values: Statistical significance (based on permutation testing)
• Normalization: Scores are normalized to enable cross-pathway comparison

Validation Rules

Input Validation

• gmtfile must be a valid enrichit database name OR accessible GMT file
• Gene names in GMT file must match Seurat object (case-sensitive)
• group_by column must exist in Seurat object metadata
• If subset_by specified, column must exist and NA values will be removed

Parameter Validation

• ntimes must be positive integer (recommended: 1000-10000)
• ncores must be positive integer (adjust based on available CPU)
• Plot types must be valid: heatmap, violin, merged_heatmap

Data Quality Validation

• Sufficient cells per group for meaningful score calculation (recommended: >10 cells)
• Gene overlap between GMT file and Seurat object (warning if too low)

Troubleshooting

Issue: Empty or all-zero pathway scores

Cause: Gene name mismatch between GMT file and Seurat object Solution:

• Check gene format in Seurat: sobj@assays$RNA@features[1:10,] (R)
• Ensure case matches: Human (UPPERCASE) vs Mouse (TitleCase)
• Verify GMT file format: name\tdescription\tgene1,gene2,gene3

Issue: Process too slow

Cause: High ntimes or insufficient ncores Solution:

# Reduce permutations for faster analysis
[ScrnaMetabolicLandscape.MetabolicPathwayActivity.envs]
ntimes = 2000  # Default is 5000

# Increase cores for parallelization
[ScrnaMetabolicLandscape.envs]
ncores = 8  # Increase based on available CPU

Issue: Out of memory errors

Cause: Large dataset with high ncores Solution:

# Reduce cores to limit parallel memory usage
[ScrnaMetabolicLandscape.envs]
ncores = 2  # Reduce from default 1 if memory issues

# Reduce permutations
[ScrnaMetabolicLandscape.MetabolicPathwayActivity.envs]
ntimes = 1000

Issue: No significant differences between groups

Cause: Groups have similar metabolic profiles or insufficient statistical power Solution:

• Increase ntimes for more robust p-value estimation
• Verify group_by column captures meaningful biological differences
• Try different pathway database (e.g., KEGG → Reactome)
• Check subset groupings if using subset_by

Issue: Subset has no cells after filtering

Cause: NA values in subset_by column or mismatched categories Solution:

# Clean metadata before analysis
[ScrnaMetabolicLandscape.envs]
mutaters = {treatment_clean = "if_else(is.na(treatment), 'unknown', treatment)"}
subset_by = "treatment_clean"

Issue: Heatmap not readable (too many pathways)

Cause: GMT file contains too many pathways for visualization Solution:

# Use smaller pathway database or custom GMT with selected pathways
gmtfile = "/path/to/core_metabolism.gmt"  # Custom curated pathways

Issue: Violin plots too crowded

Cause: Too many groups or pathways Solution:

# Generate multiple plots with subsets
[ScrnaMetabolicLandscape.MetabolicPathwayActivity.envs.plots]
"Part 1" = {
    plot_type = "violin",
    add_box = true,
    # Additional filtering options via plotthis
}
# Or use merged_heatmap for overview
"Overview" = { plot_type = "merged_heatmap" }

External References

Original Paper

Xiao, Zhengtao, Ziwei Dai, and Jason W. Locasale. "Metabolic landscape of the tumor microenvironment at single cell resolution." Nature communications 10.1 (2019): 1-12. https://www.nature.com/articles/s41467-019-11738-0

AUCell Method

AUCell: https://github.com/aertslab/AUCell - Area Under the Curve for gene set enrichment in single-cell data

Plotthis Functions

• Heatmap: https://pwwang.github.io/plotthis/reference/Heatmap.html
• ViolinPlot: https://pwwang.github.io/plotthis/reference/ViolinPlot.html

biopipen Documentation

• Metabolic pipeline: https://pwwang.github.io/biopipen/pipelines/scrna_metabolic/
• Process API: https://pwwang.github.io/biopipen/api/biopipen.ns.scrna_metabolic_landscape/

Related Skills

• ScrnaMetabolicLandscape: /skills/processes/scrnametaboliclandscape.md - Full metabolic analysis group
• MetabolicInput: /skills/processes/metabolicinput.md - Input preparation and GMT databases
• MetabolicFeatures: Pathway enrichment analysis (FGSEA-based)
• MetabolicPathwayHeterogeneity: Pathway heterogeneity analysis