Predicting the germline dependence of hematuria risk in prostate cancer radiotherapy patients

•A machine learning-based genetic model was developed to predict the risk of radiation-induced hematuria using genome-wide single nucleotide polymorphisms (SNPs) data.•The predictive model distinguished the high-risk group from the low-risk group in the validation data with the odds ratio of 2.87.•Post-hoc bioinformatics analyses identified key biological correlates that have been previously reported to be associated with the bladder and urinary tract.