Prediction of HLA-B?35∶01 susceptible components in Polygonum multiflorum Thunb.for immune-specific heterogeneous liver injury
Objective:To explore the potential HLA-B∗35∶01 susceptible components in Polygonum multiflorum,which is closely related to liver injury,by machine learning and molecular docking techniques.Methods:Small mole-cules whose adverse reactions(ADRs)are associated with HLA alleles were collected as a positive set.Small molecules whose ADRs are not currently reported to be associated with HLA alleles were collected as negative sets.The molecular descriptors were calculated by PaDEL-Descriptor software,and the binary classification model was constructed by Naïve Bayes.The model was used to predict HLA-related ADRs risk components in Polygonum multiflorum.Molecular docking was used to further analyze the binding ability as well as the binding mode with the protein crystal of HLA-B∗35∶01.Results:The average accuracy rate(ACC)of the binary classi-fication model constructed by Naïve Bayes was 0.750,and the model was used to identify 28 ADR components that may cause HLA-related ADR from 41 components of Polygonum multiflorum.Molecular docking results showed that 10 components in Polygonum multiflorum Thunb.had good binding ability with HLA-B∗35∶01.Analysis of peptides docking results showed that the binding of 2,3,5,4'-tetrahydroxy-trans-stilbene-2-O-β-glucoside,emodin-1-O-β-D-glucoside and physcion-8-O-β-D-glucopyranoside methyl ether with HLA-B∗35∶01 increased the affinity of non-HLA-B∗35∶01 presenting peptide P1(LPEPLPQGQLTAY)and P2(EECDSELEIKRY),and the three had good contacts with P1 and P2.Conclusion:In this study,machine learning and molecular docking technology predicted the potential three HLA-B∗35∶01 susceptible components in Polygonum multiflorum Radix,which may assist in the study of heterogenous drug induced liver injury components of Polygonum multiflorum Radix.
NETL
NSTL
万方数据
