Abstract:
|
Evidence suggests that the placenta has developed compensatory mechanisms to adapt to environmental exposures to preserve the developing fetus. However, the detection of such compensatory pathways has proven to be challenging. Here, we demonstrate that regression-based clustering methods (RCMs) represent a viable approach to address this problem. RCMs seek to distinguish subgroups that are driven by differing relationships between a set of features and an outcome of interest (OOI). RCMs were applied to placenta-derived RNAseq-based miRNA data collected on participants enrolled in a US-based birth cohort study, treating infant adjusted birthweight z-score as the OOI. Based on the RCM fit, we identified ~30 miRNAs whose relationship with the OOI differed between the clusters. Using putative gene target and functional analyses, we discovered that the identified miRNAs are involved in important developmental pathways. Further, we observed that cluster membership is associated with in-utero exposure to certain heavy metals, suggesting that the identified clusters may indeed reflect compensatory mechanisms developed by the placenta to ensure healthy neonatal growth and development.
|