Abstract
Despite advancements in assisted reproductive technology, recurrent implantation failure (RIF) remains a challenge. Endometrial factors, including ferroptosis and immunity, may contribute to this issue. This study integrated bioinformatics analysis and Mendelian randomization (MR) to investigate the expression and significance of DEFRGs in RIF. We intersected 484 ferroptosis-associated genes with 515 differentially expressed genes (DEGs) to identify key DEFRGs. Subsequent analyses included enrichment analysis, molecular subtype identification, machine learning model development for biomarker discovery, immune cell infiltration assessment, single-cell RNA sequencing, and MR to explore the causal relationships of selected genes with RIF. In this study, we identified 11 differentially expressed ferroptosis-related genes (DEFRGs) between RIF and healthy individuals. Cluster analysis revealed two distinct molecular subtypes with different immune profiles and DEFRG expressions. Machine learning models highlighted MUC1, GJA1 and FANCD2 as potential diagnostic biomarkers, with high accuracy in RIF prediction. Single-cell analysis further revealed the cellular localization and interactions of DEFRGs. MR suggested a protective effect of FANCD2 against RIF. Validation in RIF patients confirmed the differential expression of key DEFRGs, consistent with bioinformatics findings. This comprehensive study emphasize the significant role of DEFRGs in the pathogenesis of RIF, suggesting that modulating these genes could offer new avenues for treatment. The FANCD2 is a potential gene contributing to RIF pathogenesis through a non-classical ferroptosis-dependent pathway, providing a foundation for personalized therapeutic strategies in RIF management.