Machine learning and integrative analysis identify the common pathogenesis of azoospermia complicated with COVID-19

Abstract

Read online

BackgroundAlthough more recent evidence has indicated COVID-19 is prone to azoospermia, the common molecular mechanism of its occurrence remains to be elucidated. The aim of the present study is to further investigate the mechanism of this complication.MethodsTo discover the common differentially expressed genes (DEGs) and pathways of azoospermia and COVID-19, integrated weighted co-expression network (WGCNA), multiple machine learning analyses, and single-cell RNA-sequencing (scRNA-seq) were performed.ResultsTherefore, we screened two key network modules in the obstructive azoospermia (OA) and non-obstructive azoospermia (NOA) samples. The differentially expressed genes were mainly related to the immune system and infectious virus diseases. We then used multiple machine learning methods to detect biomarkers that differentiated OA from NOA. Enrichment analysis showed that azoospermia patients and COVID-19 patients shared a common IL-17 signaling pathway. In addition, GLO1, GPR135, DYNLL2, and EPB41L3 were identified as significant hub genes in these two diseases. Screening of two different molecular subtypes revealed that azoospermia-related genes were associated with clinicopathological characteristics of age, hospital-free-days, ventilator-free-days, charlson score, and d-dimer of patients with COVID-19 (P < 0.05). Finally, we used the Xsum method to predict potential drugs and single-cell sequencing data to further characterize whether azoospermia-related genes could validate the biological patterns of impaired spermatogenesis in cryptozoospermia patients.ConclusionOur study performs a comprehensive and integrated bioinformatics analysis of azoospermia and COVID-19. These hub genes and common pathways may provide new insights for further mechanism research.

Keywords

• azoospermia
• COVID-19
• single-cell sequencing
• machine learning
• WGCNA