New Predictive Models for Blood–Brain Barrier Permeability of Drug-like Molecules
The goals of the present study were to apply a generalized regression model and support vector machine (SVM) models with Shape Signatures descriptors, to the domain of blood–brain barrier (BBB) modeling.
Materials and Methods
The Shape Signatures method is a novel computational tool that was used to generate molecular descriptors utilized with the SVM classification technique with various BBB datasets. For comparison purposes we have created a generalized linear regression model with eight MOE descriptors and these same descriptors were also used to create SVM models.
The generalized regression model was tested on 100 molecules not in the model and resulted in a correlation r2 = 0.65. SVM models with MOE descriptors were superior to regression models, while Shape Signatures SVM models were comparable or better than those with MOE descriptors. The best 2D shape signature models had 10-fold cross validation prediction accuracy between 80–83% and leave-20%-out testing prediction accuracy between 80–82% as well as correctly predicting 84% of BBB+ compounds (n = 95) in an external database of drugs.
Our data indicate that Shape Signatures descriptors can be used with SVM and these models may have utility for predicting blood–brain barrier permeation in drug discovery.
KEY WORDSblood–brain barrier principal component analysis regression shape signatures support vector machine
absorption, distribution, metabolism and excretion
central nervous system
molecular electrostatic potential
molecular operating environment
principal component analysis
quantitative structure activity relationship
recursive feature elimination
solvent accessible surface
support vector machine
topological polar surface area
unsupervised forward selection
- 2.H. van de Waterbeemd, and E. Gifford. ADMET in silico modelling: towards prediction paradise? Nat. Rev. 2:192–204 (2003).Google Scholar
- 5.A. George. The design and molecular modeling of CNS drugs. Curr. Opin. Drug. Disc. Dev. 2:286–292 (1999).Google Scholar
- 6.K. M. Mahar Doan, J. E. Humphreys, L. O. Webster, S. A. Wring, L. J. Shampine, C. J. Serabjit-Singh, K. K. Adkison, and J. W. Polli. Passive permeability and P-glycoprotein-mediated efflux differentiate central nervous system (CNS) and non-CNS marketed drugs. J. Pharmacol. Exp. Ther. 303:1029–1037 (2002).PubMedCrossRefGoogle Scholar
- 26.L. Gomella, and S. Haist. Clinician’s pocket drug reference. McGraw-Hill, New York, 2004.Google Scholar
- 31.D. S. Chekmarev, V. Kholodovych, K. V. Balakin, Y. Ivanenkov, S. Ekins, and W. J. Welsh. Shape signatures: new descriptors for predicting cardiotoxicity in silico. Chem. Res. Toxicol., in press (2008).Google Scholar
- 32.C. Cortes, and V. Vapnik. Support vector networks. Mach. Learn. 20:273–293 (1995).Google Scholar
- 33.V. Vapnik. Statistical learning theory. Wiley, New York, 1998.Google Scholar
- 34.A. H. Fielding. Cluster and classification techniques for the biosciences. Cambridge University Press, New York, 2007.Google Scholar
- 38.C. C. Chang, and C. J. Lin. LIBSVM: A library for support vector machines, 2001.Google Scholar
- 42.H. Van de Waterbeemd, and M. Kansy. Hydrogen-bonding capacity and brain penetration. Chimia. 46:5 (1992).Google Scholar
- 50.S. Van Damme, W. Langenaeker, and P. Bultinck. Prediction of blood–brain partitioning: a model based on ab initio calculated quantum chemical descriptors. J. Mol. Graph. Model., in press (2007).Google Scholar
- 51.S. Ekins, M. J. Embrechts, C. M. Breneman, K. Jim, and J.-P. Wery. Novel applications of Kernel-partial least squares to modeling a comprehensive array of properties for drug discovery. In S. Ekins (ed.), Computational toxicology: risk assessment for pharmaceutical and environmental chemicals, Wiley, Hoboken, 2007, pp. 403–432.Google Scholar
- 52.R. Todeschini, and V. Consonni. Handbook of molecular descriptors. Wiley, Weinheim, 2000.Google Scholar