# Keratinocyte‐Associated Biomarkers Reveal Pathogenic Mechanisms in Acne

## TL;DR

This study identifies key genes in skin cells linked to acne, offering new insights into its causes and potential treatments.

## Contribution

The study introduces a novel integrative framework combining scRNA-seq, WGCNA, and machine learning to uncover acne-related keratinocyte biomarkers.

## Key findings

- Acne lesions show increased keratinocyte heterogeneity and pro-inflammatory states.
- Six keratinocyte biomarkers were identified with high diagnostic accuracy (AUC > 0.85).
- Potential drug repurposing candidates include cyclosporin A and valproic acid.

## Abstract

The cellular and molecular complexity of acne pathogenesis has hindered progress toward effective targeted therapies. While keratinocytes are known to influence skin inflammation, their precise transcriptional programs and regulatory circuitry in acne remain unclear. We developed an integrative computational framework that combines single‐cell RNA sequencing (scRNA‐seq), gene co‐expression network analysis (WGCNA), and two complementary machine learning algorithms (SVM‐RFE, LASSO) to identify disease‐relevant biomarkers. We mapped acne lesion cellular composition, reconstructed keratinocyte differentiation trajectories, and integrated miRNA–transcription factor–drug interaction networks to link molecular signatures to potential interventions. We uncovered marked keratinocyte heterogeneity and enriched late‐stage pro‐inflammatory states in acne lesions, accompanied by increased macrophage/monocyte and T cell infiltration. Six keratinocyte‐associated biomarkers (PYGL, C10orf99, C12orf75, S100A2, PI3, CARD18) were identified, achieving high diagnostic accuracy (AUC > 0.85). Functional enrichment connected these genes to cytokine and chemokine signaling, while regulatory analysis revealed upstream modulators (hsa‐let‐7b‐5p, FOXC1). Drug–gene network mapping suggested repurposing potential for cyclosporin A and valproic acid. In conclusion, our study delineates a keratinocyte‐centered molecular signature that shapes acne pathogenesis and provides potential therapeutic biomarkers.

The study workflow integrates scRNA‐seq and bulk RNA‐seq data to characterize acne pathogenesis. Step 1 identifies expanded keratinocyte populations using single‐cell analysis. Step 2 filters candidate genes via WGCNA and differential expression integration. Step 3 utilizes ensemble machine learning algorithms (Boruta, LASSO, Logistic Regression) to construct a six‐gene diagnostic nomogram. Step 4 validates these biomarkers through immune infiltration, pathway enrichment, and gene‐drug network analyses, establishing a molecular link between keratinocyte differentiation and inflammatory responses.