Construction and validation of a diagnostic model for dermatomyositis based on the LASSO algorithm

Area of Science:

• Immunology
• Bioinformatics
• Genetics

Background:

• Dermatomyositis (DM) is a prevalent autoimmune disease driven by immune responses.
• Bioinformatic analysis of immune-related genes in DM is an emerging area.
• Understanding immune characteristics is key for DM diagnosis.

Purpose of the Study:

• To construct a diagnostic model for DM using immune-related genes.
• To investigate the immune characteristics associated with DM.
• To identify potential biomarkers for DM diagnosis.

Main Methods:

• Utilized GEO datasets (GSE46239, GSE39454) for DM training and validation.
• Identified differentially expressed genes (DEGs) and intersected with immune-related genes.
• Employed LASSO regression to build a diagnostic model with 5 key genes.
• Analyzed immune cell infiltration and HLA gene expression correlations.

Main Results:

• Identified 350 DEGs and 71 immune-related DEGs.
• Developed a DM diagnostic model using ACKR1, DHX58, IRF7, ISG15, and PSMB8.
• The model demonstrated high diagnostic accuracy (AUC 0.99 training, 0.958 validation).
• Diagnostic genes correlated with M1/M2 macrophages, dendritic cells, and HLA genes.

Conclusions:

• A novel diagnostic model for DM was constructed using 5 immune-related genes.
• The model shows significant potential for aiding DM diagnosis.
• Findings highlight the association of specific immune cells and HLA genes with DM.