Roc Diagnostic Performance

ROC Diagnostic Performance

When to Use

Use this skill when you need to:

• evaluate one or more diagnostic marker genes in a case-control cohort;
• build a multivariable logistic regression diagnostic model from marker expression values;
• compare the ROC performance of the full model against individual markers.

Typical user requests:

• "Use these genes to build a diagnostic ROC model for case vs control samples."
• "Evaluate the AUC of FOXP3, CD45, and CD3E and plot all ROC curves together."
• "Run logistic regression on biomarker expression and export ROC results."

When Not to Use

Do not use this skill for:

• survival or prognostic analysis with time-to-event outcomes;
• multiclass classification tasks;
• calibration plots, nomograms, or decision-curve analysis;
• non-expression diagnostic inputs such as imaging, clinical scores, or mutation-only tables.

When to Read External Files

Usage

bashCopy code
Rscript scripts/main.R \
  --expression_file ./expression_matrix.csv \
  --group_file ./group_info.csv \
  --marker_genes FOXP3,CD45,CD3E \
  --case_group Disease \
  --output_dir ./output/ \
  --seed 42

Arguments

Input Format

Expression Matrix (expression_file)

CSV or TSV file with genes as rows and samples as columns. The first column must store unique gene identifiers.

csvCopy code
gene,Sample1,Sample2,Sample3
FOXP3,8.4,7.1,3.8
CD45,2.1,1.9,5.4
CD3E,5.8,6.2,4.0

Requirements

• File extension must be .csv, .tsv, or .txt.
• The first column must contain non-missing, unique gene identifiers.
• Remaining columns must be sample IDs.
• Selected marker genes must have numeric finite expression values across matched samples.

Group File (group_file)

CSV or TSV file with sample IDs in the first column and at least one group-label column.

csvCopy code
sample,group
Sample1,Disease
Sample2,Disease
Sample3,Control

Requirements

• File extension must be .csv, .tsv, or .txt.
• The first column must contain non-missing, unique sample IDs.
• At least one group column must be present.
• The case_group value must appear in the selected group column.
• At least 10 matched samples, 2 case samples, and 2 control samples are required.

Output Files

model_coefficients.csv

roc_auc_summary.csv

Workflow

Step 1: Validate Input

• Check that the expression matrix and group file exist and have supported formats.
• Validate unique gene identifiers and sample IDs.
• Match samples shared by both files.

Step 2: Prepare Analysis Dataset

• Keep only the requested marker genes that exist in the expression matrix.
• Merge matched expression values with group labels.
• Convert the selected case group to binary outcome labels.

Step 3: Fit Logistic Regression

• Fit a multivariable logistic regression model using the selected markers.
• Extract coefficient estimates, standard errors, p-values, and odds ratios.

Step 4: Compute ROC Performance

• Generate the ROC curve of the full logistic model.
• Generate ROC curves for each individual marker.
• Calculate AUC values for the full model and each marker.

Step 5: Save Outputs

• Save the matched analysis dataset and model bundle as .rds files.
• Save coefficient and AUC summary tables as .csv files.
• Save the combined ROC plot as a PDF.

Examples

Basic Usage

bashCopy code
Rscript scripts/main.R \
  -e expression_matrix.csv \
  -g group_info.csv \
  -m FOXP3,CD45,CD3E \
  -c Disease \
  -o ./output/

With Explicit Group Column and Custom Plot

bashCopy code
Rscript scripts/main.R \
  -e expression_matrix.csv \
  -g group_info.csv \
  -m FOXP3,CD45,CD3E \
  -c Disease \
  --group_col diagnosis \
  --plot_width 8 \
  --plot_height 6 \
  --plot_title "Biomarker ROC Comparison" \
  --legend_position topright \
  -o ./output/

With Test Data

bashCopy code
Rscript scripts/main.R \
  -e tests/data/sample_expression_matrix.csv \
  -g tests/data/sample_group_info.csv \
  -m FOXP3,CD45,CD3E \
  -c Disease \
  -o tests/expected_output/ \
  --overwrite

Error Handling

IF error persists, READ: references/troubleshooting.md

Testing

Smoke Test With Included Data

bashCopy code
Rscript scripts/main.R --help

Rscript scripts/main.R \
  -e tests/data/sample_expression_matrix.csv \
  -g tests/data/sample_group_info.csv \
  -m FOXP3,CD45,CD3E \
  -c Disease \
  -o tests/expected_output/ \
  --overwrite

Automated Smoke Test Script

bashCopy code
Rscript tests/run_smoke_test.R

Optional shell wrapper:

bashCopy code
bash tests/run_smoke_test.sh

Expected Output

textCopy code
tests/expected_output/
|-- data/analysis_data.rds
|-- data/roc_model.rds
|-- plot/roc_curve.pdf
|-- session_info.txt
|-- table/model_coefficients.csv
`-- table/roc_auc_summary.csv

References

1. Hosmer DW, Lemeshow S, Sturdivant RX (2013). Applied Logistic Regression.
2. Fawcett T (2006). An Introduction to ROC Analysis. Pattern Recognition Letters.
3. Robin X et al. (2011). pROC: an open-source package for R and S+ to analyze and compare ROC curves. BMC Bioinformatics.

For detailed algorithm, READ: references/algorithm.md

Implementation Checklist

• CLI parsing with optparse
• set.seed() for reproducibility
• requireNamespace() dependency checks
• Session info recording
• Timeout parameter exposed as CLI option
• File reading instructions in SKILL.md
• Modular script structure in scripts/
• Test data provided in tests/data/
• Error handling with SKILL_* codes
• References documented in references/

Last updated: 2026-04-17 | Version: 2.1.0