A Glycolysis Gene Methylation Prediction Model Based on Explainable Machine Learning for Alzheimer's Disease

Background: Machine learning-based disease prediction models of Alzheimer's disease(AD) are limited. This study aimed to develop a glycolysis gene methylation model for AD prediction based on machine learning, and to find biomarkers through the disease risk factor model established by Shapley Additive Explanations(SHAP). Methods: 62 elders and 60 AD patients were recruited from the affiliated hospitals of Ningbo University and assessed between 1 October 2015 to 31 December 2016. Machine learning algorithms were used to establish the model. Receiver operating characteristic curve(AUC) and Decision Curve Analysis(DCA) were used to assess the model. An AD risk factor model was established by SHAP to extract features that had important impacts on AD. Findings: The AD prediction model was developed using random forest algorithm which had the best average ROC_AUC(0.969544) in 5-fold cross-validation. The threshold probability of the model was positive in the range of 0 to 0.9875 by DCA. An AD risk factor model by SHAP had 16 important features included 8 glycolysis gene methylation indicators (GAPDHS, PKLR, PFKFB3, LDHC, DLD, ALDOC, LDHB, HK3) and 8 clinical indicators (HDL-C, TC, TG, BMI, Height, Weight, Age, Hypertension). A transcriptional factor regulatory network that played a key role in regulating 79 glycolysis-related genes was constructed. Interpretation: We developed a glycolysis gene methylation prediction model based on machine learning for AD, then established an AD risk factor model by SHAP which can be used to identify risk factors of different AD patients and provide new ideas for individualized treatment. Funding: This work was supported by National Natural Science Foundation of China(Grant no.82001155), Natural Science Foundation of Ningbo(Grant no.2019A610295) and other projects. Declaration of Interest: The authors declare that they don’t have any conflicts of interest. Ethical Approval: The study was approved by the Ethics Committee of Medical College of Ningbo University. Informed consent forms were provided by all participants or their guardians.