Objective Acute kidney transplant rejection (AKTR) remains a critical determinant of long‐term allograft survival. Emerging evidence indicates that elevated human leukocyte antigen‐F (HLA‐F) expression is linked to heightened immune injury, highlighting its potential diagnostic and therapeutic relevance. This study aimed to establish an HLA‐F–centered diagnostic model for AKTR to enhance early clinical risk stratification and to identify prognostically informative biomarkers and actionable therapeutic targets. Methods Transcriptomic datasets from kidney transplant biopsies were retrieved from Gene Expression Omnibus (GEO) and rigorously preprocessed. Differentially expressed genes (DEGs) were identified, and HLA‐F–associated dysregulated genes were used to define transcriptional changes underlying acute rejection. Functional modules linked to HLA‐F were delineated via weighted gene coexpression network analysis (WGCNA), informing the development of a 23‐gene diagnostic signature using a machine learning–based LASSO algorithm. Immune microenvironment profiling was subsequently performed by integrating machine learning–based immune deconvolution methods to quantify immune cell infiltration and assess the associations between the selected genes, immune checkpoint molecules, and tertiary lymphoid structure markers. Results HLA‐F was markedly upregulated in AKTR and exhibited strong diagnostic potential (AUC = 0.960). Integrative transcriptomic analysis identified 910 HLA‐F–associated dysregulated genes, predominantly enriched in clinically relevant pathways, including arachidonic acid metabolism, steroid hormone biosynthesis, and collecting duct acid secretion. A 23‐gene HLA‐F–centered diagnostic model was established, demonstrating strong discriminatory performance. Upstream regulatory analysis further revealed L‐tryptophan as a potential modulator of these genes through its targets IDO1 and WARS. Machine learning–based immune deconvolution showed markedly increased immune cell infiltration in the HLA‐F high‐expression group, with infiltration scores exhibiting significant positive correlations with immune checkpoint genes and tertiary lymphoid structure markers. Conclusions This study established an HLA‐F–centered diagnostic model based on 23 genes that enables quantitative HLA‐F scoring and improves risk stratification for AKTR. Beyond its diagnostic utility, the model defines a structured molecular feature set that may serve as a preliminary strategic basis for future artificial intelligence–driven drug screening efforts, pending further experimental and clinical validation.
Zhang et al. (2026) studied this question.