1、 (drug adverse reactions, ADRs) (kernel matrix dimension reduction, KMDR) , , KMDR ; ; ; (1990-) , , , , , , (20120181110051) Prediction of drug adverse reactions based on kernel matrix dimension reduction method KUANG Qifan GUO Jiali LI Yizhou LI Menglong College of Chemistry, Sichuan University;
2、Abstract Drug adverse reactions (ADRs) are quite important in drug discovery and clinical drug safety.Predicting ADRs by machine learning methods has been attracting more and more attentions.Here, a kernel matrix dimension reduction (KMDR) method was applied to infer drug ADRs and the predictive per
3、formance of this method in ADRs was investigated.The cross validation and independent tests are performed on the same standard data set with the other two reference algorithms.The results suggest that the KMDR methodcan be a promising method for ADRs prediction. Keyword cheminformatics; drug adverse
4、 reactions prediction; kernel matrix dimension reduction method; , , , , , , , , , , , 1-3 , Liu 3 , ;Huang 2 GO , , ; 2 , 1 , , Cami 1 1 , 2 , , , Logistic , ; 4 , , 5-7, , , , (kernel matrix dimension reduction, KMDR) 8, , 4 , 1 1 3 1 1.1 4 , DrugBank9 Kegg10 FAERS (FDA adverse event reporting sys
5、tem) SIDER11 2005 ( , ) , 1 , 404 461 , , , 2011 ( ) , ( ) , 1 , 2 1 - 2 1.2 - 12 SIMCOMP 13 , nd , nd d Sd, Sdij i j ; MedDRA , 14 , na , na a Sa 1.3 , Xd Xa , Xd Xa - , , , :Xd=d1, d2 nd Xa=a1, a2 na; 1 nd a Y, di aj , Yij=1, Yij=0 , , 1 nd a , di aj , di aj , 1.4 KMDR KMDR 8, , , , , 15-16, - , -
6、 , : ;S ndna ndna - ;Y ; , 1 KMDR Sd Sa - S Sd Sa ( ) K, K 1 2 n ;vi i 1 , ii i i i , 1 2 n , (2) : (3) , K i i , - S, i i : (4) , KMDR i i , S , , 086%, KMDR-KP AUC 88.1%, AUPR 34.4% , 2 , KMDR 3 , SLP-KP AUC AUPR AUC AUPR 3 4 2 3 10 4 3 10 AUPR 2.3 KMDR KMDR 1 p, p KMDR 01 6 , p=0.01, 0.05, 0.1, 0
7、.2, 0.5, 0.8, , 10 , p KMDR 3 p , KMDR , p=0.05 (KS AUPR ) , , KMDR p 0.05 KMDR , p , p , 3 p 2.4 , KMDR RLS SLP , , 4 , KP KS 2 , KMDR AUC 79.2%, AUPR20.0% , 2 avg , KMDR SLP, RLS (3 ) , KMDR 2 4 KMDR RLS SLP % 3 KMDR , 2 RLS SLP , KMDR , , , KMDR 2 , KMDR , 1CAMI A, ARNOLD A, MANZI S, et al.Predic
8、ting adverse drug events using pharmacological network modelsJ.Science Translational Medicine, 2011, 3 (114) :114-127. 2HUANG L C, WU X, CHEN J Y.Predicting adverse side effects of drugsJ.BMC Genomics, 2011, 12 (5) :S11. 3LIU M, WU Y, CHEN Y, et al.Large-scale prediction of adverse drug reactions us
9、ing chemical, biological, and phenotypic properties of drugsJ.Journal of the American Medical Informatics Association, 2012, 19 (E1) :E28-E35. 4KUANG Q, WANG M, LI R, et al.A systematic investigation of computation models for predicting adverse drug reactions (ADRs) J.PLoS One, 2014, 9 (9) :e105889.
10、 5RAYMOND R, KASHIMA H.Fast and scalable algorithms for semi-supervised link prediction on static and dynamic graphsM/BALCAZAR J L, BONCHI F, GIONIS A, et al.Machine Learning and Knowledge Discovery in Databases, 2010 (Pt III) :131-147. 6van LAARHOVEN T, NABUURS S B, MARCHIORIE.Gaussian interaction
11、profile kernels for predicting drug-target interactionJ.Bioinformatics, 2011, 27 (21) :3036-3043. 7XIA Z, WU L Y, ZHOU X, et al.Semi-supervised drug-protein interaction prediction from heterogeneous biological spacesJ.BMC Systems Biology, 2010, 4 (2) :1-16. 8KUANG Q, LI Y, WU Y, et al.A kernel matri
12、x dimension reduction method for predicting drug-target interactionJ.Chemometrics and Intelligent Laboratory Systems, 2017:doi:10.1016/j.chemolab.2017.01.016. 9WISHART D S, KNOX C, GUO A C, et al.DrugBank:a knowledgebase for drugs, drug actions and drug targetsJ.Nucleic Acids Research, 2008, 36 (Sup
13、pl 1) :D901-D906. 10KANEHISA M, GOTO S.KEGG:kyoto encyclopedia of genes and genomesJ.Nucleic Acids Research, 2000, 28 (1) :27-30. 11KUHN M, CAMPILLOS M, LETUNIC I, et al.Aside effect resource to capture phenotypic effects of drugsJ.Molecular Systems Biology, 2010, 6 (1) :343. 12WANG Y C, DENG N, CHE
14、N S, et al.Computational study of drug by integrating omics data with kernel methodsJ.Molecular Informatics, 2013, 32 (11-12) :930-941. 13HATTORI M, TANAKA N, KANEHISA M, et al.SIMCOMP/SUBCOMP:chemical structure search servers for network analysesJ.Nucleic Acids Research, 2010, 38:W652-6. 14LIN D.An
15、 information-theoretic definition of similarityC/Proceedings of the 15th International Conference on 1998:296-304. 15BROUWERS L, ISKAR M, ZELLER G, et al.Network neighbors of drug targets contribute to drug side-effect similarityJ.PLoS One, 2011, 6 (7) :e22187. 16CAMPILLOS M, KUHN M, GAVIN A C, et a
16、l.Drug target identification using side-effect similarityJ.Science, 2008, 321 (5886) :263-266. 17FAWCETT T.An introduction to ROC analysisJ.Pattern Recognition Letters, 2006, 27 (8) :861-874. 18RAGHAVAN V, BOLLMANN P, JUNG G S.A critical investigation of recall and precision as measures of retrieval system performanceJ.ACM Transactions on Information Systems (TOIS) , 1989, 7 (3) :205-229.
