KDClassifier: A urinary proteomic spectra analysis tool based on machine learning for the classification of kidney diseases

Background: We aimed to establish a novel diagnostic model for kidney diseases by combining artificial intelligence with complete mass spectrum information from urinary proteomics. Methods: enrolled 134 patients (IgA nephropathy, membranous and diabetic disease) 68 healthy participants as controls, total of 610,102 spectra their proteomic profiles. The training data set (80%) was used create using XGBoost, random forest (RF), support vector machine (SVM), neural networks (ANNs). accuracy evaluated confusion matrix test dataset (20%). also constructed receiver operating-characteristic, Lorenz, gain curves evaluate the model. Results: Compared RF, SVM, ANNs, modified XGBoost model, called Kidney Disease Classifier (KDClassifier), showed best performance. 96.03%. area under curve extreme gradient boosting (XGBoost) 0.952 (95% confidence interval, 0.9307–0.9733). Kolmogorov-Smirnov (KS) value Lorenz 0.8514. strong robustness developed Conclusions: KDClassifier achieved high thus provides potential tool classification