Endometrial receptivity profiled through transcriptomic analysis of uterine fluid extracellular vesicles using systems biology and bayesian modeling for pregnancy prediction

Identifying the optimal Window Of embryo Implantation (WOI) is important for improving pregnancy rates in Assisted Reproductive Technology (ART). During the WOI, the endometrium becomes receptive, enabling the complex communication between the embryo and endometrial tissue needed for the initiation of pregnancy. This study explores the molecular landscape of endometrial receptivity by analyzing the transcriptomic profile of Extracellular Vesicles isolated from Uterine Fluid (UF-EVs), a non-invasive alternative to traditional endometrial biopsies. RNA-sequencing of UF-EVs collected from 82 women undergoing ART with single euploid blastocyst transfer revealed 966 differentially 'expressed' genes (nominal p-value < 0.05) between women who achieved pregnancy (N = 37) and those who did not (N = 45). Patients who obtained a pregnancy showed a globally higher gene expression compared to the not-pregnant group. Weighted Gene Co-expression Network Analysis (WGCNA) clustered these differentially 'expressed' genes into four functionally relevant modules involved in key biological processes related to embryo implantation and development. A Bayesian logistic regression model, integrating gene expression modules with clinical variables, including vesicle size and history of previous miscarriages, achieved a predictive accuracy of 0.83 and an F1-score of 0.80 for pregnancy outcome prediction. This systems biology approach utilizing UF-EVs may represent an advancement over current methods that rely on endometrial transcriptomic profiles during the embryo implantation window.