Log P 是藥物開發的重要參數。藥物的親水性與其疏水性密切相關。通常 log P 的測量方法包括振盪方法，柱生成方法，高效液相色譜方法和電化學方法。搖瓶法的優點在於它簡單，具有廣泛的適用物質，並且在測量之前不需要材料結構。缺點需要是（需要至少24小時才能實現分配平衡）和大量樣品。相反，高效液相色譜的優點是快速的（每個樣品5-20分鐘）。缺點是 log P 的值是通過線性回歸確定的，因此具有相似結構的幾種化合物必須具有已知的 log P 值。基於從文獻（1995）收集的 log P 實驗數據，使用階層式支持向量回歸（HSVR）方案定量地生成計算機模型中的 log P 值。 HSVR預測在訓練集中的實驗中表現良好（n = 284, r2 = 0.87, qcv2 = 0.85, RMSE = 0.60，s = 0.36）和測試集（n = 71, r2 = 0.87, RMSE = 0.56，s = 0.31）和異常值（n = 15, r2 = 0.87, RMSE = 0.69，s = 0.35）。因此，可以斷言HSVR可以快速準確地預測藥物log P.

Log P is an important parameter for drug development. The hydrophilicity of a drug is closely related to its hydrophobicity. In general, the measurement methods of log P include an oscillation method, a column generation method, a high performance liquid chromatography method, and an electrochemical method. The advantage of the shake flask method is that it is simple, has a wide range of applicable substances, and does not require a material structure before measurement. Disadvantages need to be as long as it takes (at least 24 hours to achieve a distribution balance) and a large amount of sample. In contrast, the advantages of high performance liquid chromatography are fast (5-20 minutes per sample). The disadvantage is that the value of log P is determined by linear regression, so several compounds with similar structures must have known log P values. Based on the log P experimental data collected from the literature (1995), the log P values in the computer model were quantitatively generated using a hierarchical support vector regression (HSVR) scheme. The HSVR predictions performed well in experiments with those in the training set (n = 284, r2 = 0.87, qcv2 = 0.85, RMSE = 0.60，s = 0.36) and the test set (n = 71, r2 = 0.87, RMSE = 0.56，s = 0.31) and outliers (n = 15, r2 = 0.87, RMSE = 0.69，s = 0.35). Therefore, it can be asserted that HSVR can predict drug log P quickly and accurately.

謝辭 iii
摘要 v
ABSTRACT vii
TABLE OF CONTENTS ix
FIGURE xi
TABLE xiii
ABBREVIATIONS xv
1. Introduction 1
2. Material and method 3
2-1. Data collection 3
2-2. Dataset selection 3
2-3. Molecular descriptors 3
2-4. Descriptor selection 4
2-5. Support Vector Machine 5
2-6. Support Vector Regression 5
2-7. Hierarchical Support Vector Regression 6
2-8. Predictive evaluation 6
3. Result 11
3-1. Dataset selection 11
3-2. SVR models 11
3-3. HSVR 12
3-4. Predictive evaluations 14
3-5. Mock test 15
3-6. Comparison with previous models 15
4. Discussion 17
5. Conclusion 19
6. References 21

Andrés, Axel, et al.
2015 Setup and validation of shake-flask procedures for the determination of partition coefficients (logD) from low drug amounts. Eur. J. Pharm. Sci. 76:181-191.
Atkinson, Hartley C., and Evan J. Begg
1988 Relationship between Human Milk Lipid-Ultrafiltrate and Octanol-Water Partition Coefficients. J. Pharm. Sci. 77(9):796-798.
Austin, Rupert P., Andrew M. Davis, and Carol N. Manners
1995 Partitioning of ionizing molecules between aqueous buffers and phospholipid vesicles. J. Pharm. Sci. 84(10):1180-1183.
Avery, Mitchell A., et al.
1995 Structure-Activity Relationships of the Antimalarial Agent Artemisinin. 2. Effect of Heteroatom Substitution at O-11: Synthesis and Bioassay of N-Alkyl-11-aza-9-desmethylartemisinins. J. Med. Chem. 38(26):5038-5044.
Bayat, Z., and J. Movaffagh
2010 Evaluation of the 1-octanol/water partition coefficient of nucleoside analogs via free energy estimated in quantum chemical calculations. Russ. J. Phys. Chem. A 84(13):2293-2299.
Besler, Brent H., Kenneth M. Merz Jr., and Peter A. Kollman
1990 Atomic charges derived from semiempirical methods. J. Comput. Chem. 11(4):431-439.
Breiman, Leo, and Philip Spector
1992 Submodel Selection and Evaluation in Regression. The X-Random Case. Int. Stat. Rev. 60(3):291-319.
Burges, Christopher J.C. %J Data Mining, and Knowledge Discovery
1998 A Tutorial on Support Vector Machines for Pattern Recognition. Data Min. Knowl. Discov. 2(2):121-167.
Cappelli, Claudia Ileana, Emilio Benfenati, and Josep Cester
2015 Evaluation of QSAR models for predicting the partition coefficient (log P) of chemicals under the REACH regulation. Environmental Research 143, Part A:26-32.
Chen, Rena Zhanglei, Walter Z. Tang, and Mika Sillanpää
2016 Prediction of Log P of Halogenated Alkanes by Their ELUMO and Number of Chlorine and Carbon. Environ. Process. 3(1):73-91.
Cheng, Tiejun, et al.
2007 Computation of Octanol−Water Partition Coefficients by Guiding an Additive Model with Knowledge. Journal of Chemical Information and Modeling 47(6):2140-2148.
Chiou, Cary T., David W. Schmedding, and Milton Manes
2005 Improved Prediction of Octanol−Water Partition Coefficients from Liquid−Solute Water Solubilities and Molar Volumes. Environ. Sci. Technol. 39(22):8840-8846.
Chirico, Nicola, and Paola Gramatica
2011 Real External Predictivity of QSAR Models: How To Evaluate It? Comparison of Different Validation Criteria and Proposal of Using the Concordance Correlation Coefficient. J. Chem Inf. Model. 51(9):2320-2335.
—
2012 Real External Predictivity of QSAR Models. Part 2. New Intercomparable Thresholds for Different Validation Criteria and the Need for Scatter Plot Inspection. J. Chem. Inf. Model. 52(8):2044-2058.
Consonni, Viviana, Davide Ballabio, and Roberto Todeschini
2009 Comments on the Definition of the Q2 Parameter for QSAR Validation. J. Chem Inf. Model. 49(7):1669-1678.
Cortes, Corinna, and Vladimir %J Machine Learning Vapnik
1995 Support-vector networks. Mach. Learn. 20(3):273-297.
Danielsson, Lars-Go¨ran, and Yu-Hui Zhang
1996 Methods for determiningn-octanol-water partition constants. TrAC Trends. Anal. Chem. 15(4):188-196.
de Melo, Eduardo Borges , and Márcia Miguel Castro Ferreira
2009 Nonequivalent Effects of Diverse LogP Algorithms in Three QSAR Studies. QSAR Comb. Sci. 28(10):1156-1165.
Dearden, John C., and Gaynor M. Bresnen
1988 The Measurement of Partition Coefficients. Quant. Struct.-Act. Relat. 7(3):133-144.
Debnath, Asim Kumar, et al.
1991 Structure-activity relationship of mutagenic aromatic and heteroaromatic nitro compounds. Correlation with molecular orbital energies and hydrophobicity. J. Med. Chem. 34(2):786-797.
Ghose, Arup K., and Gordon M. Crippen
1986 Atomic Physicochemical Parameters for Three-Dimensional Structure-Directed Quantitative Structure-Activity Relationships I. Partition Coefficients as a Measure of Hydrophobicity. J. Comput. Chem. 7(4):565.
Gnanadesikan, R., and J. R. Kettenring
1972 Robust Estimates, Residuals, and Outlier Detection with Multiresponse Data. Biometrics 28(1):81-124.
Golbraikh, Alexander, et al.
2003 Rational selection of training and test sets for the development of validated QSAR models. J. Comput.-Aided Mol. Des. 17(2):241-253.
Gramatica, Paola, and Alessandro Sangion
2016 A Historical Excursus on the Statistical Validation Parameters for QSAR Models: A Clarification Concerning Metrics and Terminology. J. Chem. Inf. Model. 56(6):1127-1131.
Hansch, C., A. Leo, and D. Hoekman
1995 Exploring QSAR: Hydrophobic, Electronic, and Steric Constants. Volume 2. Washington, DC: American Chemical Society.
Ivanciuc, Ovidiu
2007 Applications of Support Vector Machines in Chemistry. In Reviews in Computational Chemistry. Pp. 291-400.
Kettaneh, Nouna, Anders Berglund, and Svante Wold
2005 PCA and PLS with very large data sets. Computational Statistics & Data Analysis 48(1):69-85.
King, R. D., et al.
1996 Structure-activity relationships derived by machine learning: the use of atoms and their bond connectivities to predict mutagenicity by inductive logic programming. Proc. Natl. Acad. Sci. 93(1):438.
Klein, Markus, Lothar Erdinger, and Gernot Boche
2000 From mutagenic to non-mutagenic nitroarenes: effect of bulky alkyl substituents on the mutagenic activity of nitroaromatics in Salmonella typhimurium: Part II. Substituents far away from the nitro group. Mutation Research/Genetic Toxicology and Environmental Mutagenesis 467(1):69-82.
Klopman, Gilles, and Linda D. Iroff
1981 Calculation of partition coefficients by the charge density method. J. Comput. Chem. 2(2):157-160.
Kril, Michael B., and Ho-Leung Fung
1990 Influence of hydrophobicity on the ion exchange selectivity coefficients for aromatic amines. J. Pharm. Sci. 79(5):440-443.
Leo, Albert, Corwin Hansch, and David Elkins
1971 Partition coefficients and their uses. Chem. Rev. 71(6):525-616.
Leong, Max K., Yen-Ming Chen, and Tzu-Hsien Chen
2009 Prediction of human cytochrome P450 2B6-substrate interactions using hierarchical support vector regression approach. J Comput Chem 30(12):1899-1909.
Leong, Max K., et al.
2010 Predicting Mutagenicity of Aromatic Amines by Various Machine Learning Approaches. Toxicol. Sci. 116(2):498-513.
Moriguchi, Ikuo, et al.
1992 Simple Method of Calculating Octanol/Water Partition Coefficient. CHEM. PHARM. BULL. 40(1):127-130.
Ojha, Probir Kumar, et al.
2011 Further exploring rm2 metrics for validation of QSPR models. Chemometrics Intell. Lab. Syst. 107(1):194-205.
Ranadive, Sunanda A., Andrew X. Chen, and Abu T. M. %J Pharmaceutical Research Serajuddin
1992 Relative Lipophilicities and Structural-Pharmacological Considerations of Various Angiotensin-Converting Enzyme (ACE) Inhibitors. Pharm. Res. 9(11):1480-1486.
Recanatini, M.
1992 Partition and distribution coefficients of aryloxypropanolamine β-adrenoceptor antagonists. J. Pharm. Pharmacol. 44(1):68-70.
Ritchie, Timothy J., and Simon J. F. Macdonald
2009 The impact of aromatic ring count on compound developability – are too many aromatic rings a liability in drug design? Drug Discovery Today 14(21):1011-1020.
Rostad, Colleen E., Wilfred E. Pereira, and Marc F. Hult
1985 Partitioning studies of coal-tar constituents in a two-phase contaminated ground-water system. Chemosphere 14(8):1023-1036.
Roy, Kunal, et al.
2012 Comparative Studies on Some Metrics for External Validation of QSPR Models. J. Chem Inf. Model. 52(2):396-408.
Sangster, James
1989 Octanol‐Water Partition Coefficients of Simple Organic Compounds. J. Phys. Chem. Ref. Data 18(3):1111-1229.
Schüürmann, Gerrit, et al.
2008 External Validation and Prediction Employing the Predictive Squared Correlation Coefficient — Test Set Activity Mean vs Training Set Activity Mean. J. Chem Inf. Model. 48(11):2140-2145.
Taft, R. W., et al.
1985 Solubility Properties in Polymers and Biological Media 5: An Analysis of the Physicochemical Properties Which Influence Octanol–Water Partition Coefficients of Aliphatic and Aromatic Solutes. J. Pharm. Sci. 74(8):807-814.
Topliss, John G., and Robert P. Edwards
1979 Chance factors in studies of quantitative structure-activity relationships. J. Med. Chem. 22(10):1238-1244.
Tropsha, Alexander
2010 Best Practices for QSAR Model Development, Validation, and Exploitation. Mol. Inf. 29(6‐7):476-488.
Tropsha, Alexander, Paola Gramatica, and Vijay K. Gombar
2003 The Importance of Being Earnest: Validation is the Absolute Essential for Successful Application and Interpretation of QSPR Models. 22(1):69-77.
Valkó, Klára
2004 Application of high-performance liquid chromatography based measurements of lipophilicity to model biological distribution. Journal of Chromatography A 1037(1):299-310.
Zou, J. W., et al.
2002 A Quantitative Structure-Property Relationship Analysis of logP for Disubstituted Benzenes. J. Phys. Chem. A 106(47):11550-11557.