tf-target-gene-regulatory-network

name	tf-target-gene-regulatory-network
description	Use when analyzing transcription factor (TF) regulatory networks using Dorothea database. Input gene list, identify regulating transcription factors, generate TF-Target network visualization. For: transcription factor enrichment analysis, gene regulatory network research.
license	MIT
skill-author	AIPOCH

Transcription Factor (TF) Regulatory Network Analysis

When to Use

• Use this skill when you have a human or mouse gene list and want to identify upstream TFs from the Dorothea database.
• Use it when you need a ready-to-export TF-target network table plus a publication-ready PDF network plot.
• Use it for reproducible CLI execution with saved session information and optional local Dorothea .rds databases.

When Not to Use

• Do not use this skill for differential expression, pathway enrichment, cell type annotation, or survival analysis.
• Do not use it to infer causal direction beyond curated Dorothea TF-target relationships.
• Do not use it when your input genes are aliases or mixed-species symbols that have not been normalized first.

Input Validation

This skill accepts: a human or mouse gene list (HGNC symbols for human, first-letter-uppercase for mouse) for TF regulatory network analysis using Dorothea.

If the user's request does not involve identifying upstream transcription factors from a gene list — for example, asking to run differential expression, pathway enrichment, cell type annotation, or multi-omics integration — do not proceed with the workflow. Instead respond:

"tf-target-gene-regulatory-network is designed to identify upstream transcription factors from a gene list using the Dorothea database and generate a TF-target network visualization. Your request appears to be outside this scope. Please provide a gene list for TF regulatory analysis, or use a more appropriate tool for your task."

Entry Point

• Primary CLI entry point: scripts/main.R
• Canonical visualization values: English tokens fr, curve, diamond, triangle, square

When to Read External Files

Situation	File to Read	Purpose
Need algorithm details	references/algorithm.md	Statistical methods, Dorothea database, network analysis algorithms
Need to run analysis	scripts/main.R	Execute: Rscript scripts/main.R --gene ... --species ...
Encounter errors	references/troubleshooting.md	Common errors and solutions
Need CLI examples	references/cli-guide.md	Detailed CLI usage examples
Need test data	tests/data/	Sample gene lists for testing

---

Installation

R Package Dependencies

# CRAN packages
install.packages(c("optparse", "dplyr", "openxlsx", "tidyverse", "tidygraph", "ggraph", "showtext"))

# Bioconductor packages (optional if using local database files)
if (!require("BiocManager", quietly = TRUE))
    install.packages("BiocManager")
BiocManager::install("dorothea")

Local Database Files (Recommended)

For faster analysis and offline use, generate local database files:

Rscript database/database-get.R

This creates database/dorothea_hs.rds (human) and database/dorothea_mm.rds (mouse) in the skill root directory.

Verification

Rscript scripts/main.R --help

Usage

Rscript scripts/main.R \
  --gene "TP53,MYC,EGFR" \
  --species human \
  --output_dir ./TF_Result \
  --seed 42

---

Arguments

Note: Either --gene or --gene_file must be provided (at least one is required).

Short	Long	Type	Default	Description
-g	--gene	character	NULL	Comma-separated gene list (e.g., "TP53,MYC,EGFR") — required if --gene_file not provided
-f	--gene_file	character	NULL	File with gene names (txt or csv, one per line or comma-separated) — required if --gene not provided
-s	--species	character	human	Species: human or mouse
-o	--output_dir	character	TF_Result	Output directory name
	--db_path	character	NULL	Local .rds database file path. If not specified, auto-searches default paths
-d	--dir	character	NULL	Working root directory (advanced)
	--seed	integer	42	Random seed for reproducibility
	--title	character	""	Main plot title

Visualization Parameters: For complete list (plot dimensions, colors, labels, layout, edge styles), see references/visualization-parameters.md. Canonical values are English tokens: fr (force-directed layout), curve (curved edges), diamond / triangle / square (node shapes).

---

Input Format

Gene List Input

Two ways to provide input genes:

1. Command line: --gene "TP53,MYC,EGFR" (comma-separated)
2. File input: --gene_file genes.txt (one gene per line or comma-separated)

Gene Naming Convention

• Human genes: All uppercase symbols (e.g., TP53, MYC, EGFR)
• Mouse genes: First letter uppercase only (e.g., Tp53, Myc, Egfr)
• Use official gene symbols, not aliases
• Case-sensitive for species matching

Species Support

• human: Human genes (Homo sapiens)
• mouse: Mouse genes (Mus musculus)

Performance Guidance

For large gene lists (> 500 genes), Dorothea database queries may take several minutes. Use a local .rds database (--db_path) for substantially faster lookups. There is currently no --timeout_seconds parameter; monitor progress with verbose logging if available.

---

Output Files

File	Description
TF_Network_Plot.pdf	TF-target network visualization
tf_network.xlsx	Network data (edges and nodes worksheets)
TF_Target_Filtered_Core_<species>.xlsx	Complete TF-target relationships table
session_info.txt	R session and package version info
tf.Rdata	R environment data

Outputs are organized under TF_Result/ in data/, plot/, and table/ subdirectories.

---

Workflow

Step 1: Load Database

• Priority search for local database files (.rds format) in: --db_path, getwd()/database/, script_dir/database/, dirname(script_dir)/database/
• If no local file found, load from Dorothea R package
• Filter for high-confidence interactions (confidence levels A, B, C)

Step 2: Identify Regulating TFs

• Match input genes against target genes in Dorothea database
• Extract transcription factors regulating these targets
• Compute TF frequency (number of targets regulated)

Step 3: Generate Network Data

• Create edge list (TF → Target relationships)
• Create node list with types (TF or Target)
• Save to Excel format for downstream analysis

Step 4: Visualize Network

• Generate network graph using tidygraph and ggraph
• Apply visual customization (layout, colors, shapes)
• Save as PDF publication-ready figure

---

Methods

Dorothea Database

• Curated database of TF-target interactions
• Confidence levels: A (high), B (medium), C (low)
• Includes human and mouse data
• Source: https://github.com/saezlab/dorothea

Network Analysis

• Graph construction: TF-target relationships as directed edges
• Layout algorithms: Multiple options including fr (force-directed), circle, grid, sphere
• Visual customization: Full control over colors, shapes, sizes

Local Database Feature

• Option to use pre-saved .rds files for faster analysis and offline use
• Priority search paths for local database files (see Step 1 above)
• Fallback to Dorothea R package if no local file found

---

Examples

Basic Usage (Human Genes)

Rscript scripts/main.R \
  -g "TP53,MYC,EGFR" \
  -s human \
  -o ./TF_Result

File Input

Rscript scripts/main.R \
  -f gene_list.txt \
  -s human \
  -o ./TF_Result

Mouse Genes

Rscript scripts/main.R \
  -g "Tp53,Myc,Egfr" \
  -s mouse \
  -o ./Mouse_TF_Result

Custom Styling

Rscript scripts/main.R \
  -g "PTPRC,FOXP3,CD4" \
  -s human \
  --style_layout "fr" \
  --style_line "curve" \
  --point_shape "diamond,triangle" \
  --line_color "#E64B35" \
  --title "Immune TF Network" \
  -o ./Custom_Plot

Using Local Database

Rscript scripts/main.R \
  -g "TP53,MYC,EGFR" \
  -s human \
  --db_path database/dorothea_hs.rds \
  -o ./LocalDB_Result

---

Error Handling

Common Errors

Error Code	Cause	Solution
SKILL_FILE_NOT_FOUND	Input gene file does not exist	Check file path and permissions
SKILL_NO_INPUT_GENES	Empty gene list or file	Provide genes using --gene or --gene_file
SKILL_INVALID_SPECIES	Species not human or mouse	Use human or mouse only
SKILL_INVALID_PARAMETER	Invalid layout, shape, line, or legend value	Use supported values shown by Rscript scripts/main.R --help
SKILL_EMPTY_RESULTS	No TF-target relationships found for input genes	Check gene symbols and species; try broadening confidence levels
SKILL_DEPENDENCY_MISSING	Missing dplyr, dorothea, tidygraph, etc.	Install missing packages (see Installation section)

Exit Status Codes

Code	Meaning
0	Success
1	Execution error (see error code for details)
2	SKILL_EMPTY_RESULTS — no TF-target matches found for the input genes

IF error persists, READ: references/troubleshooting.md

---

Testing

Test with Sample Data

# Check help
Rscript scripts/main.R --help

# Run with sample human genes
Rscript scripts/main.R \
  -g "TP53,MYC,EGFR" \
  -s human \
  -o tests/output_human/

# Run with sample mouse genes
Rscript scripts/main.R \
  -g "Tp53,Myc,Egfr" \
  -s mouse \
  -o tests/output_mouse/

After running, verify tests/output_human/plot/TF_Network_Plot.pdf and tests/output_human/table/tf_network.xlsx exist and are non-empty.

Reference Files

File	Purpose
references/algorithm.md	Statistical methods and Dorothea database details
references/troubleshooting.md	Common errors and solutions
references/cli-guide.md	CLI usage examples
references/visualization-parameters.md	Complete visualization parameter list