Multi-omics analyses and machine learning prediction of oviductal responses in the presence of gametes and embryos

Ryan M Finnerty,Daniel J Carulli,Akshata Hedge,Yanli Wang,Frimpong Boadu,Sarayut Winuthayanon,Jianlin Cheng,Wipawee Winuthayanon

DOI: https://doi.org/10.1101/2024.06.13.598905

2024-06-15

Abstract:The oviduct is the site of fertilization and preimplantation embryo development in mammals. Evidence suggests that gametes alter oviductal gene expression. To delineate the adaptive interactions between the oviduct and gamete/embryo, we performed a multi-omics characterization of oviductal tissues utilizing bulk RNA-sequencing (RNA-seq), single-cell RNA-sequencing (scRNA-seq), and proteomics collected from distal and proximal at various stages after mating in mice. We observed robust region-specific transcriptional signatures. Specifically, the presence of sperm induces genes involved in pro-inflammatory responses in the proximal region at 0.5 days post-coitus (dpc). Genes involved in inflammatory responses were produced specifically by secretory epithelial cells in the oviduct. At 1.5 and 2.5 dpc, genes involved in pyruvate and glycolysis were enriched in the proximal region, potentially providing metabolic support for developing embryos. Abundant proteins in the oviductal fluid were differentially observed between naturally fertilized and superovulated samples. RNA-seq data were used to identify transcription factors predicted to influence protein abundance in the proteomic data via a novel machine learning model based on transformers of integrating transcriptomics and proteomics data. The transformers identified influential transcription factors and correlated predictive protein expressions in alignment with the in vivo-derived data. In conclusion, our multi-omics characterization and subsequent in vivo confirmation of proteins/RNAs indicate that the oviduct is adaptive and responsive to the presence of sperm and embryos in a spatiotemporal manner.

Cell Biology