Machine Learning Analysis On MicroRNA Expression Data In Diagnosis Of Wind-phlegm-stasis Syndrome Of Ischemic Stroke

Objective: To establish micro RNA models for the diagnosis of windphlegm-stasis syndrome of ischemic stroke(IS)by using machine learning methods;and to explore the potential of micro RNA being a diagnostic marker of wind-phlegm-stasis syndrome of IS,so as to provide a molecular basis for Traditional Chinese Medicine syndrome dignosis of stroke.Methods: The expression level of micro RNAs in the peripheral blood of 50 patients diagnosed with wind-phlegm-stasis syndrome of IS and 50 matched controls were assessed through real-time polymerase chain reaction(q RT-PCR).Machine learning algorithms,including artificial neural network,random forest,e Xtreme Gradient Boosting,and support vector machine(SVM)were employed via R 3.6.3 software to establish diagnostic models for wind-phlegm-stasis syndrome of IS.Results: The wind-phlegm-stasis syndrome group has significantly increased expression level of miR-19a(P<0.001),miR-148a(P<0.001),miR-320d(P=0.003),and miR-342-3p(P<0.001)compared with the control group.Mi R-148 a,miR-342-3p,miR-19 a,and miR-320 d yielded areas under the receiver operating characteristic curve(AUC)of 0.872,0.844,0.721,and 0.673,respectively,with 0.740,0.940,0.740,and 0.840 sensitivity and 0.920,0.640,0.600,and 0.440 specificity,respectively.The expression level of miR-148 a and miR-342-3p in peripheral blood of patients was significantly negatively correlated with thrombin time(TT).In addition,patients with higher level of miR-19 a,miR-148 a or miR-320 d had a lower level of TT.Patients with higher level of miR-342-3p had a higher level of both triglyceride and low-density lipoprotein.Model miR-148 a + miR-342-3p + miR-19 a had the best predictive value when analyzed via SVM algorithm with AUC,sensitivity,and specificity values of 0.958,0.937,and 0.889,respectively.Conclusions:(1)MiR-19 a,miR-148 a,miR-320 d,and miR-342-3p had the ability to diagnose wind-phlegm-stasis syndrome of IS,while they might attribute to the disease by affecting coagulation function and blood lipid.(2)The diagnostic value of the expression data increased after being intergrated via machine learning algoryrhms.The diagnostic value of the combination of miR-148 a,miR-342-3p,and miR-19 a through SVM algorithm can serve as a feasible approach to promote the diagnosis of wind-phlegm-stasis syndrome of IS.