A Bayesian Hierarchical Modeling Study on Factors Influencing Male Fetal Y-Chromosome Concentration

This study develops a Bayesian hierarchical model to analyze factors influencing male fetal Y-chromosome concentration (Y%), crucial for Non-Invasive Prenatal Testing (NIPT) accuracy. Initially, Spearman’s correlation analysis identified 13 significant indicators, including gestational week (positive correlation) and maternal BMI/weight (negative correlation). Given Y%’s non-negative, right-skewed, and long-tailed distribution, distribution fitting confirmed its alignment with a Rice distribution. A Bayesian hierarchical model was constructed incorporating these indicators as fixed effects, along with individual maternal random effects to account for heterogeneity. Parameters were estimated via MCMC sampling (NUTS algorithm, 300,000 iterations), with all chains converging effectively (R