# Identification and validation of lactate-related gene signatures in endometriosis for clinical evaluation and immune characterization by WGCNA and machine learning

## TL;DR

This study identifies key lactate-related genes in endometriosis, offering potential for early diagnosis and immune mechanism insights.

## Contribution

Novel lactate-related gene signatures and a diagnostic model for endometriosis using WGCNA and machine learning.

## Key findings

- Three core lactate-related biomarkers (BPGM, DHFR, SLC25A13) were identified for endometriosis.
- A nomogram model showed strong diagnostic performance for endometriosis.
- Immune dysregulation linked to lactate-related genes was revealed in endometriosis.

## Abstract

Endometriosis is a common benign gynecologic disease in women of reproductive age, and its manifestations remarkably decrease quality of life. Lactate, as a metabolite, exerts prominent effects across a wide range of biological processes. The objective of this research is to explore the clinical value and immune features of lactate-related genes in endometriosis and contribute novel strategies for guiding the clinical management of patients with endometriosis.

We first conducted a differential expression analysis to identify the differentially expressed genes (DEGs) in the training set. By integrating the critical module genes from weighted gene co-expression network analysis (WGCNA) and lactate-related genes (LRGs), we preliminarily screened lactate-related differentially expressed genes (LR-DEGs). Machine learning algorithms, single-cell datasets, and clinical samples were used to further identify and validate core LR-DEGs. Subsequently, we evaluated the diagnostic value of the model constructed from core LR-DEGs for endometriosis and explored the biological functions of these genes. Additionally, we conducted immune-related analysis in endometriosis and identified small molecule compounds targeting core LR-DEGs.

In this study, 22 candidate genes were identified by intersecting 2,318 DEGs and 2,177 key module genes with 357 LRGs. This list was further refined using three machine learning algorithms, resulting in three primary lactate-related biomarkers: BPGM, DHFR, and SLC25A13. A nomogram model constructed from core LR-DEGs demonstrated outstanding diagnostic performance in identifying patients with endometriosis. Immune-related analysis revealed significant associations between hub LR-DEGs and cellular immune dysregulation in endometriosis. Additionally, a gene–small molecule compound regulatory network was established to guide potential treatment strategies.

Taken together, our study established a robust relationship between lactate metabolism-related genes and endometriosis, with the model promising to enable the early diagnosis of endometriosis, contribute to the excavation of the immune molecular mechanisms of endometriosis, and support the discovery of potential targets for therapy in the metabolism of endometriosis.