Mining biomarkers for infertility-associated conditions. Studies on polycystic ovary syndrome and recurrent implantation failure through microbiome and AI-based approaches

Female infertility accounts for about 35% of all infertility cases and is often caused by conditions like polycystic ovary syndrome (PCOS) and recurrent implantation failure (RIF). Women with PCOS have irregular menstrual cycles, often with infrequent or absent ovulation, and are more likely to experience mood disorders (MDs) than women without PCOS. RIF refers to repeated failed attempts at in vitro fertilization (IVF) after transfer with good quality embryos. Both PCOS and RIF are associated with increased systemic and uterine inflammation and endometrial dysfunction during implantation, which can raise the risk of infertility. Despite extensive research, the exact causes and optimal treatments for these conditions are still unclear.
This thesis dived into identifying biomarkers for PCOS and RIF using microbiome and artificial intelligence (AI)-based analysis. Scientists have been exploring the potential of microbiome, the collection of all microbes living in and on the human body, to better understand and treat diseases. This thesis examined the vaginal and endometrial microbiome of women with PCOS and identified thirteen microbial features associated with PCOS. The thesis also found that women with PCOS who experienced MDs have distinct gut microbiome profiles compared to those without MDs and revealed correlations between the abundance of specific microbes and common clinical traits of both PCOS and MDs. Additionally, the thesis utilized an AI model to study endometrial development and inflammation-related traits in women with PCOS and RIF. The thesis revealed differences in the occurrence of endometrial CD138+ plasma cells—used as markers for chronic endometrial inflammation—based on PCOS phenotype. However, the endometrial receptivity status in RIF women did not affect either endometrial gland development or the occurrence of CD138+ plasma cells.
This thesis highlights the potential of microbiome and AI-based approaches in uncovering new insights into PCOS and RIF, paving the way for better understanding and treatment of these conditions.